A nonlinear iterative filtering based on local statistics and anisotropic diffusion is introduced. Local statistics determines the diffusion coefficient at each iteration step. Anisotropic diffusion can be seen as estimates a piecewise smooth image from the noisy input image in the experimental section, our results are shown to suppress noise with preserving the edges. Therefore, it enhances the image and improves performance.

This paper proposes the multiple objects focusing algorithm. Given multiple objects at different distances from a camera, we assume that one object is well-focused and the others are out-of-focused. The proposed auto-focusing algorithm is summarized as follows: (i) detects edges from an input image, (ⅱ) estimates the radius of PSF (Point Spread Function) across the edge, (ⅲ) gather edge points having same radius of PSF, (ⅳ) segments the image into regions with the same radius of PSF, and (ⅴ) restores the each segmented region using the corresponding PSF.

In block-based lossy video compression, severe quantization causes discontinuities along block boundaries so that annoying blocking artifacts are visible in decoded video imases. These blocking artifacts significantly decrease the subjective image quality. In order to reduce the blocking artifacts in decoded images, many algorithms have been proposed However studies on so called, corner outliers, have been very limited. Corner outliers make image edges look disconnected from those of neighboring blocks at cross block boundary. In order to solve this problem, we propose a corner outlier detection and compensation algorithm as post-processing in spatial domain The experiment results show that the proposed method provides much improved subjective image quality.

This paper proposes a color matching 3D look-up table simplifying the complex color matching procedure between a monitor and a mobile display device. In order to perform color matching, it is necessary to process color of image in the device independent color space like CIEXYZ or CIELAB. We improved the S-curve model to have smaller characterization error than tolerance error. Also, as a result of the experiments, we concluded that the color matching look-up table with 64(4$\times$4$\times$4) is the smallest size allowing characterization error to be acceptable.

This paper proposes the method of visual resolution enhancement to render a color image on a delta-structured display. The proposed method adopted a subpixel rendering method to reduce a color fringe error caused by delta- structured display and to improve visual resolution

In this work. we propose an algorithm for detecting and compensating dynamic false contours in plasma display panels (PDPs). First, we detect the candidate pixels, which are likely to be corrupted by false contours, and merge those pixels into several regions. Second, we estimate the motion vectors of the selected regions. Finally, based on the motion information. we modify the luminance values of the pixels in the regions to alleviate the effects of false contours. Simulation results demonstrate that the proposed algorithm efficiently reduces dynamic false contours at low computational complexity.

In this paper a new approach of digitalized ‘Takbon’ is introduced. By image signal processing, the letters which were written on stones can be deciphered. Epigraphic letter is detected by digital image device, digital camera. The two dimensional digital image is preprocessed because of sensor noise and detective turbulence. Color image is transformed into grey level. The letter image is analyzed in time/frequency domain. By the resultant analysis data decisive functions are calculated. Signal Processing techniques, such as scaling, clipping, digital negative, high/low filter, morphology and so on, provide algorithms that can extract letter from stones.

In this paper, we describe a method for separating text lines in handwritten Korean address images. The most remarkable feature of the proposed method is to use a modified projection technique. named a partial projection technique. A projection based text line separation method which projects the whole address image in horizontal direction to find split points for text line separation cannot avoid failing separation in case of images with a little skew or overlap between vertically neighboring text lines. To overcome this problem, we have introduced a partial projection technique which splits an address image into a few partial address images to be equal width and then project them each horizontally. The experiment done with 989 handwritten Korean address images extracted from live mails shows the superiority of the proposed method. The correct text-line separation rate fir the testing images was about 91.5％.

In this paper, I propose an automatic background extraction method and continuous background updating technique. Because there is a movement of a vehicle and a change of a background is feeble, the area moving through the time axis is looked for and a background and a vehicle image is divided. A way to give dynamically the threshold which divides the image frame into a vehicle image and the background in a space is enforced. Through the repetition of the above-mentioned process, the background pictorial image is gained. Using the karlman filter technique, the update is done so that a background image can obey a climate situation and an environmental change in day and night. A background image processed algorithm is better than the existent one. Through simulation, the feasibility of the algorithm has been verified.

In this paper, we propose a background segmentation method in low resolution color image. A segmentation algorithm is based on color and edge information. In edge image, adaptive and local thresholds are applied to suppress paint boundaries. Through our experiments, the proposed algorithm efficiently segments background from objects.

In general, speech coding methods are classified into the following three categories: the waveform coding, the source coding and the hybrid coding. Fast speaking is possible to encode with a few information compared with slow speaking rate. In case of speaking rate, low frequency band is more important than high frequency band while listening. Speech vocoding technique is developing to way with low bit rate and complexity and high sound quality. the CELP type of vocoder support very good sound quality with low bit rate but these vocoders don't consider about the speaking rate. When we consider speaking rate and encode the frame depending on the speaking rate, the bit rate is able to reduce the bit rate than the conventional vocoder. We propose the technique to estimate the speaking rate and applied PSOLA technique in case of the frame of slow speaking rate. As a result of simulation bit rate can be reduced about 300 bps.

In this paper, we propose new speech feature parameter using NMf(Non-Negative Matrix Factorization). NMF can represent multi-dimensional data based on effective dimensional reduction through matrix factorization under the non-negativity constraint, and reduced data present parts-based features of input data. In this paper, we verify about usefulness of NMF algorithm for speech feature extraction applying feature parameter that is got using NMF in Mel-scaled filter bank output. According to recognition experiment result, we could confirm that proposal feature parameter is superior in recognition performance than MFCC(mel frequency cepstral coefficient) that is used generally.

In this paper, we propose a new method which generates the trajectory of the mouth shape for the characters by the user inputs. It is based on the character at a basis syllable and can be suitable to the mouth shape generation. In this paper, we understand the principle of the Korean language creation and find the similarity for the form of the mouth shape and select it as a basic syllable. We also consider the articulation of this phoneme for it and create a new mouth shape trajectory and apply at face of an 3D avatar.

In this paper, we propose a new excitation enhancement technique to improve the speech quality of low bit rate speech coders. The proposed technique is based on a harmonic model and it is employed only in the decoding process of speech coders without any additional bits. We develop the procedure of harmonic model parameters estimation and harmonic generation. and apply the technique to a current state of the art low bit rate speech coder, ITU-T G.729 Annex D. Also its performance is measured by using the ITU-T P.862 PESQ score and compared to those of the phase dispersion filter and the long-term postfilter applied to the decoded excitation. It is shown that the proposed excitation enhancement technique can improve the quality of decoded speech and provide better quality for male speech than other techniques.

In this paper, we implement PAMD(Positive Answering Machine Detection) for discrimination human and ARS. We are used Grunt detection, Glitch Noise detection and Tone detection for PAMD. It distinguishes voice signals from ring-back tone and glitch noise respectively. And as a second step, it judges whether human responses or ARS responses after integrating pattern changes like initial response period, the number of voice data, each time of voice data period and glitch noise. The accuracy is about 9375 in ASR and about 98％ in Mobile phone.

In this paper, a speech recognition system using a speaker defendant algorithm is implemented on the PC. Results are loaded on a LDM display system that employs Intel StrongArm SA-1110. This research has completed so that this speech recognition system may correct its shortcomings. Sometimes a former system is operated by similar speech, not a same one. To input a vocalization is processed two times to solve mentioned defects. When references are creating, variable start-point and end-point are given to make efficient references. This references and new references are changed into feature parameter, LPC and MFCC. DTW is excuted using feature parameter. This security system will give user permission under fore execution have same result.

본 논문에서는 범용 DSP를 이용한 잡음환경에 강인한 음성인식 시스템을 구현하였다. 구현된 시스템은 TI사의 범용 DSP인 TMS320C32를 이용하였고, 실시간 음성 입력을 위한 음성 Codec과 외부 인터페이스를 확장하여 인식결과를 출력하도록 구성하였다. 또한, 기존의 음성 인식 시스템에 사용한 파라메터에 대한 고찰과 ICA를 이용하여 잡음 환경에 강인한 음성 특징 파라메터를 제안하고 성능 비교 실험을 하였다. 제안된 ICA 파라메터를 적용하여 음성인식 시스템을 구현하였다. 그리고, 독립적으로 동작 가능한 음성인식 시스템의 응용 예로 무선자동차에 적용시켜 실험했다.

본 논문에서는 MPEG-2 BC와 AAC의 복호화 과정 중 함성 필터링 과정의 알고리듬을 분석하여 공동된 구조로 연산을 수행한 수 있는 광용 합성 필터 구조에 대하여 논하였다. 제안된 공용 합성 필터 구조는 Regressive 구조를 이용하여 MPEG-2 BC와 AAC의 복호화를 효과적으로 공용 수행하도록 하였다. 제안한 구조는 FFT를 사용할 경우에 필요한 전처리 및 후처리 과정을 고려해주지 않아도 되고 복소수 연산이 아닌 실수연산이 되어 하드웨어 구조가 단순하게 된다. 또한 MPEG-2 AAC의 다양한 윈도우 변환에도 안정적으로 연산되는 구조임을 확인하였다.

본 논문에서는 디지털방송과 인터넷의 융합에 따른 MPEG-2/4/7 방송 및 인터넷 콘텐츠를 비롯한 게임등과 같은 다양한 멀티미디어 서비스를 제공하기 위한 차세대 지능형 고품질 홈 엔터테인먼트 시스템 Platform 개발에서 사용될 MPEG-4 오디오를 개발한다. 인터넷 상에서의 스트리밍 서비스를 위해서는 저 전송률과 고 품질의 비디오/오디오 알고리즘이 필요하다. 이러한 서비스를 제공하기 위하여 MPEG-4 오디오는 음성에서 고품질의 다중 채널의 오디오까지, 그리고 자연음(Natural Sound)에서 합성음에 이르기까지 다양한 알고리즘을 제공한다. 본 논문에서는 지능형 고품질 미디어 에이전트 시스템에 적합한 MPEG-4 AAC, MPEG-1 Layer-3인 MP3, G.723.1을 구현하고, 이 시스템에 알맞은 7㎑ 대역폭을 가지는 광대역(Wideband) 음성신호를 16kbps로 압축하는 음성 압축기를 제안 및 개발한다.

In this paper, we propose a new hardware architecture for integer transform, quantizer operation of a new video coding standard H.264/JVT. We describe the algorithm to derive hardware architecture emphasizing the importance of area for low cost and low power consumption. The proposed architecture has been verified by PCI-interfaced emulation board using APEX-II Altera FPGA and also by ASIC synthesis using Samsung 0.18 ${\mu}{\textrm}{m}$ CMOS cell library. The ASIC synthesis result shows that the proposed hardware can operate at 100 MHz, processing more than 1, 300 QCIF video frames per second. The hardware is going to be used as a core module when implementing a complete H.264 video encoder/decoder ASIC for real-time multimedia application.

This paper presents a variable block size motion estimation (ME) algorithm and hardware architectures dedicated to H.264/AVC. Proposed ME architecture can achieve real-time processing for 720$\times$480@30Hz with search range of [-64, +63] in the horizontal and [-32, +31] in the vertical direction at integer-pel accuracy and upto 7 reference frames at the operating frequency of 54MHz.

H.264/AVC is the upcoming video coding standard of ITU-T H.264 and ISO MPEG-4 AVC. The new standard can achieve a significant improvement up to 50％ in compression ratio compared to MPEG-4 advanced simple profile. In this paper, we propose the novel intra prediction scheme to speed up intra prediction process in H.264/AVC decoder and show the hardware architecture for it. The proposed scheme uses the concurrent processing of the 4$\times$4 intra prediction, which is based on that some 4$\times$4 block pairs in a 16$\times$16 luma block can be processed concurrently. The proposed scheme can reduce intra prediction time by 33 ％.

To provide high-quality media streaming service over the best-effort Internet, an efficient streaming system with a network-adaptive transport is essential. By adopting FEC-based adaptive transport and by matching it to the priority of media packets, in this paper, an enhanced MPEG-4 streaming system is implemented. The implemented adaptive transport can complement the existing MPEG-4 streaming server with components such as elementary stream provider, sync layer and DMIF layer. Sync layer packets of each MPEG-4 elementary stream are prioritized and then selectively dropped at the sender to match the currently available bandwidth. Also, with the proposed adaptive transport module, the level of FEC protection is dynamically adjusted based on the underlying network condition. With preliminary verification over the emulated network testbed, it is shown that the proposed implementation can mitigate the impact of network fluctuation and thus improve the quality of streaming.

Computing environments of Embedded Systems are different from those of desktop computers so that they have resource constraints such as CPU processing, memory capacity, power, and etc.. Thus, when a desktop S/W is ported into embedded systems, optimization should be seriously considered. In this paper, we investigate several S/W optimization techniques to be considered for porting H.263 encoder into a high performance DSP, TMS320C6711. Through experiments, it is found that optimization techniques employed can make a big performance improvement.

An oscillatory neural network circuit has been designed and fabricated in an 0.5 ${\mu}{\textrm}{m}$ double poly CMOS technology. The proposed oscillatory neural network consists of 3 neural oscillator cells with excitatory synapses and a neural oscillator cell with inhibitory synapse. Simulations of a network of oscillators demonstrate cooperative computation. Measurements of the fabricated chip in condition of $\pm$ 2.5 V power supply is shown.

본 논문은 LTJ 적응필터의 실용적 구현에 관한 연구이다. 음성코덱(codec)을 사용하는 응용분야에서는 코덱 복호화단의 LPC 계수정보를 얻을 수 있으므로 이를 반사계수로 변환하여 사용하므로서 반사계수 적응에 소용되는 계산량을 감소시킬 수 있으며, 코덱에서는 프레임 또는 서브프레임 단위로 LPC 계수를 적응시키므로 시변 변환 영역 적응필터에 해당하는 LTJ 적응필터의 필터 계수 보상에 필요한 계산량을 감소시킬 수 있다. 실제 음성신호를 사용하여 제안된 실용적 구현 방법의 타당성을 검증하였다.

In this paper, an industrial electric endoscope is implemented. Most industrial endoscope that used in domestic are imported and the equipments are divided into several parts. So these endoscopes spend a lot of time, labor and inspection cost on inspection process. This research is accomplished to solve above mentioned weak points and to improve previous inspection systems into embedded systems.

In this paper we implemented a field-sequential stereoscopic endoscope system that can generate stereoscopic images with different perspective depth using LCD stutter. Re stereoscopic image is generated from stereoscopic adapter that has LCD shutter. We have compared the stereoscopic depth of a field-sequential stereoscopic endoscope system with that of the conventional endoscope system. And the implemented system is verified by evaluation the field-sequential stereoscopic image on a Monitor. This system will be use to medical instruments in time.

This paper presents an efficient huffman decoding method based on the multiple branch technique. In the proposed search method, the internal node which does not contain a leaf node are removed for decrease the searching time and the memory consumption. The proposed search method gives 44％ of improved in searching time and 34％ of decreased in memory requirement compared to the binary search method.

In this paper, we propose two methods for complexity reduction of intra prediction in H.264/AVC. One is skipping of intra prediction using inter prediction cost at current macroblock in current P picture, average of intra prediction cost in previous I picture, and average of inter prediction cost in previous P picture. The other is skipping of intra 16$\times$16 prediction using intra 4$\times$4 prediction cost and modes. As a result, complexity of intra prediction in P picture and that of intra 16$\times$16 prediction in intra prediction macroblock can be reduced by about 80~99％ and 50~93％, respectively.

Fixed GOP allocation in MPEG-2 cannot cope with scene change and amount of motion, which results in degradation picture quality. By finding suitable N and M and allocating dynamic GOP, the improvement of picture quality can be achieved. In this paper, N and M are determined by scene change detection and estimation of amount of motion using color histogram per each macroblock. The simulation results show that the average PSNR is improved, especially around the shot boundaries.

In this paper, we propose a low bit-rate embedded video coding scheme with 3-D block partition in the wavelet domain. The proposed video coding scheme includes multi-level three dimensional dyadic wavelet decomposition, raster scanning within each subband, partitioning of blocks, and adaptive arithmetic entropy coding. Although the proposed video coding scheme is quite simple, it produces bit-streams with good features, including SNR scalability from the embedded nature. Experimental results demonstrate that the proposed video coding scheme is quite competitive to other good wavelet-based video coders in the literature.

In this paper, we propose a new construction algorithm for the reversible variable-length code (RVLC) using a simplified average length function of the optimal Huffman code. RVLC is introduced as one of the error resilience tools in H.263+ and MPEG-4 owing to its error-correcting capability. The proposed algorithm demonstrates an improved performance in terms of the average codeword length over the existing HVLC algorithms.

가장 최근의 동영상 압축 표준인 H.264는 매크로블록의 최적 모드를 결정하기 위하여 총 7가지 모드를 사용하여 움직임 예측을 수행하기 때문에 부호화 효율과 수행 성능 면에서 기존의 방식에 비해 좋은 결과를 보이지만, 움직임 예측 과정에서 많은 계산 시간을 필요로 한다는 것이 커다란 단점으로 지적되고 있다. 따라서, 본 논문에서는 H.264에서 가변 블륵 단위의 움직임 예측 시 인접한 블록과의 상관성을 분석하여 다음 움직임 예측 모드의 사용 여부를 결정하는 블록 병합 알고리즘을 HEXBS(Hexagon-based Search) 고속 움직임 탐색 알고리즘에 적용하여 움직임 예측 시간을 효율적으로 절약하는 방법을 제안한다.

움직임 벡터의 상관도, 움직임 벡터의 분포특성, 블록 정합 오류의 특성은 탐색 패턴과 탐색 방법을 결정하는 중요한 요소이다. 일반적으로 움직임 벡터는 주로 탐색영역의 가운데를 중심으로 수평 흑은 수직축에 주로 분포한다. 또한 탐색 영역 내의 정합 오류 값의 분포를 보면 움직임 벡터의 분포와 비슷한 형태로 정합 오류의 값들이 수평 혹은 수직 방향으로 최소 정합 오류 값의 위치로 단조 감소해 나간다. 본 논문에서는 이러한 블록 정합 오류의 특성을 이용한 새로운 탐색 방법을 제안하며 주변 블록의 움직임 벡터의 상관도를 이용하여 초기 탐색 지점을 선택하는 고속 움직임 추정 알고리즘을 제안한다. 또한 모의실험을 통하여 기존의 여러 움직임 추정 알고리즘과 비교하여 PSNR 의 감소는 거의 없으면서 매크로블록당 평균 탐색포인트와 수행 시간의 향상을 얻을 수 있음을 확인한다.

This paper presents a method to find motion vectors that are closer to true motion with noisy images for simultaneous noise reduction and do-interlacing. The proposed method requires four interlaced field images: one noisy field image and three field images from which noise is already removed. The validation of motion provides accurate motion vectors and allows us to utilize them even in very noisy environment. The validated motion vectors are first used for the noise reduction, buffered and used later for the noise reduction and de -interlacing.

This paper proposes a video tracking method for a deformable moving object using an active contour model. In order to decide the convergent directions of the contour points automatically, a new energy function based on a frame difference map and an updating rules of the frame difference map are presented. Experimental results on a set of synthetic and real image sequences showed that the proposed method can fully track a speedy deformable object while extracting the boundary of the object exactly in every frame.

In this paper. we propose a simple and effective video scene change detection algorithm using a 3-D DCT. The 3-D DCT that we employ is a 2$\times$2$\times$2 DCT has simple computations composed only of adding and shifting operations. The simple average values of multiresolution represented video using the 2$\times$2$\times$2 DCT are used as a detection feature vector.

This paper proposes a hierarchical approach to active shape model using wavelet transform. The proposed algorithm allows us to use both global shape characteristics and finer details for model deformation. The statistical properties of the wavelet transform of a deformable model are analyzed by principal component analysis and used as priors in the contour's deformation.

We propose stereoscopic video coding method using the syntax of MAC(Multiple auxiliary component) that was added to the MPEC-4 visual version 2 in order to describe the transparency of the video object. We also define the novel MAC's semantics in MPEG-4 that should support the proposed coding method. The major difference between the existing coding method and the proposed coding method is the addition of the residual texture coding. The proposed coding method assigns disparity map and residual texture to 3 components of MAC: one component for disparity map and the rest 2 components fer the luminance and chrominance data of the residual texture, respectively. The performance of the proposed method is evaluated in terms of PSNR by computer simulations.

In this paper, we propose the method of stereo images composition using adaptive dense disparity estimation. For the correct composition of stereo image and 3D virtual object, we need correct marker position and depth information. The existing algorithms use position information of markers in stereo images for calculating depth of calibration object. But this depth information may be wrong in case of inaccurate marker tracking. Moreover in occlusion region, we can't know depth of 3D object, so we can't composite stereo images and 3D virtual object. In these reasons, the proposed algorithm uses adaptive dense disparity estimation for calculation of depth. The adaptive dense disparity estimation is the algorithm that use pixel-based disparity estimation and the search range is limited around calibration object.

This paper proposes a stereo image insertion technique on DCT coefficients using the embedded method developed in the digital watermarking in due consideration of compatibility with conventional 2D system such JPEG, and MPEG. In conventional transmission method users with conventional digital TV cannot watch the transmitted 3D image sequence as 2D image, because of affectivity of conventional 3D image compression. To give an answer, in this paper, DCT coefficients are changed according to its disparity on YCbCr channels. Our method can insert stereo images into a conventional image compression method based on DCT.

This paper proposes a new stereo matching algorithm using both the divide-and-conquer method and the DSI(Disparity Space Image) technique. Firstly, we find salient feature points on the each scanline of the left image and find the corresponding feature point at the right image. Then the problem of a scanline is divided into several subproblems. By this way, matching of the subintervals is implemented by using the DSI technique. The DSI technique for stereo matching process is a very efficient solution to find matches and occlusions simultaneously and it is very speedy. In addition, we apply three occluding patterns to process occluded regions, as a result, we reduce mismatches at the disparity discontinuity.

A vision sensor should be calibrated prior to infer a Euclidian shape reconstruction. A point to point calibration. also referred to as a hard calibration, estimates calibration parameters by means of a set of 3D to 2D point pairs. We proposed a new method for determining a set of 3D to 2D pairs for the structured light hard calibration. It is simply determined based on epipolar geometry between camera image plane and projector plane, and a projector calibrating grid pattern. The projector calibration is divided two stages; world 3D data acquisition Stage and corresponding 2D data acquisition stage. After 3D data points are derived using cross ratio, corresponding 2D point in the projector plane can be determined by the fundamental matrix and horizontal grid ID of a projector calibrating pattern. Euclidian reconstruction can be achieved by linear triangulation. and experimental results from simulation are presented.

In this paper, we propose an extrinsic calibration method for a multi-view camera to get an optimal pose in 3D space. Conventional calibration algorithms do not guarantee the calibration accuracy at a mid/long distance because pixel errors increase as the distance between camera and pattern goes far. To compensate for the calibration errors, firstly, we apply the Tsai's algorithm to each lens so that we obtain initial extrinsic parameters Then, we estimate extrinsic parameters by using distance vectors obtained from structural cues of a multi-view camera. After we get the estimated extrinsic parameters of each lens, we carry out a non-linear optimization using the relationship between camera coordinate and world coordinate iteratively. The optimal camera parameters can be used in generating 3D panoramic virtual environment and supporting AR applications.

The recovery of 3D scene structure from multiple views has been long one of the central problems in computer vision. This paper presents a new projective reconstruction method based on factorization for un-calibrated image sequences. The proposed algorithm provides an effective measure to construct frame groups by using various information between frames. The experimental results show that the proposed method can reconstruct a more precise 3D structure than the precious methods such as the merging method.

Camera calibration is the method of finding the relationship between world coordinates and image coordinates with respect to camera coordinate. It is represented by PPM(Perspective Projection Mateix) compactly. And the PPM is effected by positions, intervals and distribution of control points. This paper presents to the relationship between PPMs and control points by experiment.

In this paper, we propose a new 3D coding method for heterogeneous systems over enhanced Access Grid (e-AG) with 3D display using spatio-temporal scalability. The proposed encoder produces four bit-streams: one base layer and enhancement layer l, 2 and 3. The base layer represents a video sequence for left eye with lower spatial resolution. An enhancement layer l provides additional bit-stream needed for reproduction of frames produced in base layer with full resolution. Similarly, the enhancement layer 2 represents a video sequence for right eye with lower spatial resolution and an enhancement layer 3 provides additional bit-stream needed for reproduction of its reference pictures with full resolution. In this system, temporal resolution reduction is obtained by dropping B-frames in the receiver according to network condition. The receiver system can select the spatial and temporal resolution of video sequence with its display condition by properly combining bit-streams.

This study is focused on proposing a creative system that can display 3D hologram on the real-time basis. This method applies 3D display on volume hologram based on CGH. The process of implementing the system consists of two stages of fringe pattern recording for passive component that includes information on hologram, and irradiating object beam. Distinguished from an existing electronic holographic display system, this system is free from the process of a huge calculation that is necessary to compose CGH for a real-time 3D display.

In this paper, we propose a digital watermarking method using the spread spectrum techniques in DCT domain. Watermark is embedded in the middle band frequencies of 8$\times$8 DCT block. Therefore, we improve image quality and robustness. Experiment results show that the proposed watermark is robust against various attack.

This paper presents a wavelet-based digital watermarking scheme for color images. We insert watermarks in the DWI domain using spread-spectrum correlation-based watermarking in luminance component of the color image. The watermark, two pseudorandom patterns, is inserted by modifying the wavelet coefficients at subband HL2, LH2, HH2 according to message bits. To detect watermark, we calculate the correlation between the watermarked image and pseudorandom patterns. As compared with correlations, the watermark detector determines embedding message. Experimental results show that the scheme is not only invisible but also robust to various attacks such as JPEG compression, noise addition and so on.

Recently, TFT-LCD is widely used of medicine machine on the display devices. However, the display precision of TFT-LCD is 8 bits instead of 10 bits of CRT display. If the medical image have more than 8 bits, we must requantize the medical image. We propose an efficient method to reduce medical image from 10 bits into 8 bits by employing human visual perception. The proposed method shows good performance for the medical image display.

The purpose of this paper is to develop a semi -automated system for medical image processing with which tissues or organs from medical images can be segmented and classified by people who have basic knowledge of image processing. In addition, the proposed medical image processing system is independent on types of human tissues or images. In this paper, a new semi-automated image processing system with essential image processing functions for medical images is introduced

In this paper, we present an efficient algorithm to detect microcalcifications ROI (Regions of Interest) in digital mammograms using Linear filters. To efficiently detect microcalcifications ROI, we used three sequential processes; preprocessing for breast area detection, modified multilevel thresholding, ROI selection using mean filter and linear filters.

In this paper, we propose a method to search the evidence such as a knife found in the crime scene based on GHT from an image database Such objects like knives are simitar in shape. The proposed method utilizes the small shape differences among objects as much as possible to distinguish an object from similar shaped objects. It consists of the GHT based candidate generation and top-down candidate verification. For the fast generation of the candidate 1ist, the GHT operation is performed un the down sampled edge list. The test results show that it can retrieve the correct object even with a pan of object in reasonable time.

We implement a system for 3D structure reconstruction from multiple 2D images. It uses geometric primitives such as box, wedge, pyramid, etc, each having translation, rotation, and scale parameters. Primitives are marked on input images with GUI (Graphic User Interface). Lines made by projection of primitives onto an image correspond to marked line segments of the image. Error function is defined by disparity between them and is minimized by downhill simplex method. By assigning relationship between models, the number of parameters to solve can be decreased and the resultant models become more accurate To share variables among other models also reduces computational complexity. Experiments using real images have shown that the proposed method successfully reconstructs 3D structure.

Nowadays image processing is very useful for some field of traffic applications. The one reason is we can construct the system in a low price, the other is the improvement of hardware processing power, it can be more fast to processing the data. In this study, I propose the traffic monitoring system that implement on the embedded system environment. The whole system consists of two main part, one is host controller board, the other is image processing board. The part of host controller board take charge of control the total system, interface of external environment. and OSD(On screen display). The part of image processing board takes charge of image input and output using video encoder and decoder, image classification and memory control of using FPGA, control of mouse signal. And finally, fer stable operation of host controller board, uC/OS-II operating system is ported on the board.

This parer proposes the six color separation using th color difference and granularity. Conventional method using the color difference increases the graininess in the bright region due to the usage of the cyan or magenta. To reduce the graininess in the bright region, we proposed the six color separation minimizing the graininess within the tolerance of the co]or difference. Initially, granularity is calculated based on the standard deviation of the lightness value and chrominance of the SCIELAB space and is applied to the six color separation using the color difference. Proposed six color separation using the color difference and granularity reduces the graininess in the bright region and obtains the smooth tone.

본 논문의 목표는 디지털방송과 인터넷의 융합에 따른 MPEG-2/4/7 방송 및 인터넷 콘텐츠을 비롯한 게임등과 같은 다양한 멀티미디어 서비스를 제공하기 위한 차세대 지능형 고품질 홈 엔터테인먼트 시스템 Platform 개발이다. 디지털 방송과 데이터방송 수신이 가능한 Set-Top Box기능, 수신된 방송의 저장 및 재생이 가능한 PDR 기능, MPEG-2 형식을 MPEG-4 형식으로 변환하는 Transcoding 기능, VOD 서비스를 제공하기 위한 Streaming 기능 등을 지원할 수 있는 시스템의 구조를 설계하였다. 이러한 지능형 고품질 서비스를 지원하기 일해 고성능 시스템이 필요하다. 시스템 제어를 위한 CPU 로는 PMC-Sierra사의 MIPS Architecture에 기반을 둔 RM5231 을 채택하고, MPEC-4 Decoding, BIFS Presentation Engine과 Streaming 서비스와 MPEC-7 Metadata Generator/Parser 을 위해 ARM Architecture에 기반을 둔 Intel80200 Processor 를 채택하였다. 또한, 디지털방송을 위한 MPEC-2 Decoder Chip 인 Teraloglc 사의 TL811 System Controller 와 TL851 Graphics& Display Processor 를 채택하였다. 개발된 시스템을 테스트하기 위하여 DVB-MHP Server와 MPEG-4 IP Streaming Server 를 구축하여 디지털 방송과 Streaming 서비스를 테스트하였다.

TCP is taking over 95％ among the Internet traffics. Recently the demands of multimedia services in the Internet has been increasing. These multimedia services mostly need real-time deliverly, and then RTP has been a de facto to transmission protocol for these real-time multimedia services. RTP uses UDP as its underlying transport protocol, and thus it does not support any rate and congestion control. Thus, for fair use of the Internet bandwidth with TCP traffics. RTP also needs a rate control. One constraint of RTP is that the feedback information(delivered by, RTP's twin protocol, RTCP) is recommended to be sent no less than 5 seconds. In this paper, we propose a TCP-friendly RTP rate control which use only RTCP feedback information at every 5 seconds. The experiment results show that our proposed algorithm works. But, it is found that we need more time to test the effects of parameters and policies of the algorithms, which will be reported later.

A efficient P2P (peer to peer) networking for sharing multimedia contents is proposed to overcome the limitations of unstructured approaches. The proposed approach is attempting to organize the participating P2P nodes by modifying a n-dimensional cartesian-coordinate space DHT (distributed hash table), CAN (content addressable network). The network identifiers (e.g., network prefix of IP address) of participating nodes are mapped into the CAN virtual coordinate space (in 2-d) and nodes with similar identifiers are grouped into the same zone. The proposed scheme is expected to show some level of concentration reflecting the network identifiers. Network simulator-based evaluation is performed to verify the effectiveness of the proposed scheme.

In this paper, we propose a new progressive image transmission system including the image coding scheme that efficiently uses the relationship between the properties of a spatial image and its wavelet transform. Firstly, an original image is decomposed into several layers by the wavelet transform, and simultaneously decomposed into 2"x2" blocks. Each image is classified into two image types according to the standard deviations of its blocks. And then each block is categorized into two regions by different thresholds according to the image types, i.e., significant activity region (SAR) and insignificant activity region (IAR). Simulation results show that the proposed coding method has better performance than the EZW and SPIHT in terms of image quality and transmitted bit-rate. In addition, it can be applied to the applications requiring the progressive image transmission.nsmission.

The DMB recently standardized in Korea is a Eureka-147 DAB(Digital Audio Broadcasting)-based standard which is able to provide multimedia services including moving pictures, still images, text and etc. That has the structure to add the MPEG media codec and the MPEG system, namely, video-multiplexer to the DAB system. In this paper, we analysis the video-multiplexer of the DMB standard and propose a new multiplexer, namely. M4GM(MPEG-4 General Mux) included in the DMB vido-multiplexer for the performance improvement with respect to the transmission efficiency and the expansible functions. In addition, we simulate the two video-multiplexers and then compare and estimate their performance entirely.

This paper presents a robust error concealment method for burst-packet-loss networks. The proposed error concealment algorithm can reduce the computational complexities of the existing error concealment methods. Moreover, experimental results show that the proposed method produces the better video quality than the conventional boundary matching algorithm.

Terrestrial DMB(Digital Multimedia Broadcasting) system is the standard that offers multimedia broadcasting services at mobile environment and is based on Eureka-147 DAB(Digital Audio Broadcasting) for transmission method. Also DMB provides the error protection method of convolution coding. In this paper, we study on the effective error resilience coding of MPEG-4 video stream over DMB system. To accomplish error resilience, we first partition one data into several data using the data partitioning, and we control the coding rate of the convolution coding according to the importance of the partitioned data. In this algorithm, we suggest and analyze the efficient rate control algorithm considering convolution code rate.

Terrestrial DMB(Digital Multimedia Broadcasting) system is the standard that offers multimedia broadcasting services at mobile environment and is based on Eureka-147 DAB(Digital Audio Broadcasting) for transmission method. Also DMB provides the error protection method of convolution coding. In this paper, we study on the effective error resilience coding of MPEG-4 video stream over DMB system. To accomplish error resilience, we first partition one data into several data using the data partitioning, and we control the coding rate of the convolution coding according to the importance of the partitioned data. In this algorithm, we suggest and analyze the efficient rate control algorithm considering convolution code rate.

This paper presents a novel signal processing method for optimal beamforming of smart antenna system in cdma2000 1X mobile communication environments. This method utilizes characteristics of the reverse pilot channel of cdma2000 1X mobile communication systems, and applies them to improve the performance of an adaptive algorithm, which is used to a smart antenna system for beamforming.

In recent years, one of the main trends of GIS is to construct and provide realistic and human-perceptible information of spatial object using video data. The rapid progress of wireless mobile technology makes the multimedia source using video data possible on mobile environment. Instead of traditional map-based service, video-based location services, that provide video data as location context, would be of interest in connection with various mobile applications requiring location service, such as telematics, LBS, etc. In this paper, we present the location service system using geo-referenced video on wireless mobile environment.

The enhancement of fingerprint image is necessary to improve the performance of fingerprint recognition. The enhancement of fingerprint image with Gabor Filter(GF) is widely used. However GF has the weakness such as long processing time and the sensitivity to ridge frequency. To overcome these weaknesses, we propose a Modified Anisotropic Gaussian Filter(MAGF) which is modified from Anisotropic Filter proposed by S. Greenburg's(SAF). This proposed MAGF can reduce the calculation time of ridge frequency and improve the weakness of sensitivity to ridge frequency. We also explained that MAGF is better than others mathematically and experimentally.

In this paper, fingerprint minutiae mosaicking algorithm using minutiae of fingerprint is proposed. First, minutiae map is generated from minutiae of fingerprint and minutiae constellation is generated from fingerprint minutiae map. Minutiae constellation is constellation-shaped structure generated from Voronoi Diagram and Delaunay Triangulation using information of minutiae. Secondly, common region is detected by similarity of minutiae constellation of fingerprint minutiae map and minutiae map of individual fingerprint image is composed. Consequently composite minutiae map by mosaicking of fingerprint minutiae improve the performance of the fingerprint matching system.

This raper propose a matching algorithm using voronoi diagram for rotation and translation invariant fingerprint identification. The proposed algorithm extracts geometrical structures that ate derived from voronoi diagram of a fingerprint image. Then distances and angles are extracted from the geometrical structure and saved indexing form for fingerprint matching. Experimental results show that the proposed algorithm invariant to fingerprint rotation and translation requirements and matching time.

The postprocessing of fingerprint image are widely used to eliminate the false minutiae that caused by skeletonization. This paper presents a new postprocessing algorithm of the skeletonized fingerprint image using SOFM. The proposed postprocessing method showed the good performance for eliminating the spurious minutiae.

기존 홍채 인식 시스템의 접안식 영상 획득상 불편한 사항을 해결하고 인식의 정확도를 높이기 위해서는 원격으로 영상을 획득할 수 있어야 하며, 홍채의 경계선을 정확하게 검출할 수 있어야 한다. 또한 기존 홍채 영역 검출 방법의 문제점인 홍채를 원으로 가정하는 방식을 개선할 필요성이 있다. 따라서 본 논문에서는 조명에 의한 glint 정보와 intensity gradient를 이용하여 홍채의 경계를 산출하였으며, 아울러 스테레오스코픽 카메라를 이용하여 홍채 경계의 3차원 좌표를 획득함으로써, 카메라를 기준으로 하는 홍채의 주시각을 찾아 홍채의 원형 변환에 활용하도록 하였다.

This paper proposes a new measurement method of face region size for real-time serveillance systems. The proposed method consists of three steps. In the first step, it detects global face area based on the block-based difference images. In the second step, it measures the face region size using face color information. In the third step, it estimates the face position in the next input image using the trajectory of face regions. The experimental results have shown that the proposed algorithm measures the face size within 20％ relative error on average, which is a]towable for most surveillance systems.

In this paper, we present real-time, accurate face region detection and tracking technique for an intelligent surveillance system. It is very important to obtain the high-resolution images, which enables accurate identification of an object-of-interest. Conventional surveillance or security systems, however, usually provide poor image quality because they use one or more fixed cameras and keep recording scenes without any clue. We implemented a real-time surveillance system that tracks a moving person using pan-tilt-zoom (PTZ) cameras. While tracking, the region-of-interest (ROI) can be obtained by using a low-pass filter and background subtraction. Color information in the ROI is updated to extract features for optimal tracking and zooming. The experiment with real human faces showed highly acceptable results in the sense of both accuracy and computational efficiency.

This paper presents a novel system that analyzes and recognizes a gait based on shape context on silhouette images. The main functions of the system consist of three steps: First, the system extracts the silhouette images from galt image sequence by performing a simple pre-processing and acquires the AGM(Averaged Gait Map) by averaging them. Next. it computes the cross-correlation between the AGMs. Finally, it classifies the AGMs based on the cross-correlation using nearest neighborhood classification. The proposed system uses two cues to classify a gait: One corresponds to biometric shape cue such as body height width. and body-part proportions. The other corresponds to gait cue such as stride length and amount of arm swing. Perceptionally, the biometric cues are sailent on the double support (both legs spread and touching the ground) while the gait cues on the midstance. Through a variety of experiments, it is proved that the property of a gait is mainly influenced by gait cues than biometric cues.

In this paper, we tried to develop a vowel training assistant method using vowel formant statistics. Formant statistics were obtained from PBW set consists of 452 words from 8 persons. Then, we calculated distance from input formants to each center of vowel formant space. Based on the distance, directions to correct the speaker's manner of articulation, i.e. position of jaw and tongue.

In this paper a new OSD (On-Screen Display) development system for digital devices is presented. The proposed system is an example of the ubiquitous systems because it can be used to co-develop an OSD tn any time at any place if the web can be accessed. In UI, OSD is an important pan of digital devices. Especially, OSD implementations for multiple languages are essential to export digital devices and make them user-friendly. The proposed system consists of four components: font editor, menu editor. R/C simulator, and screen editor. It provides all the functions to develop an OSD and the result can be downloaded to a digital device. The proposed system was used to develop an OSD for 31 countries and the lead time of development was reduced from 52.7 days to 20 days.

The recognition limit of brightness difference in the visual system of human has not been studied system-atically. In this paper, endeavor to calculate surround adapted Weber-Fechner fractions based on the crispening effect was added to the traditional Weber-Fechner fractions. It is found that surround adapted fractions have reduced to 1/3 of the traditional fractions. As compared with Breitmeyer's experiments, the presented result is a reasonable one. It can be used as some guide to design the digital display system when a designer needs to decide bit count of digital signal in considering of the limit of brightness level, and as the inspection tool of display manufacturing of brightness smear, defect, and so on.

An PIAGP generator for TFT-LCD production has been proposed in this paper. The proposed generator was implemented using careful control of timing controller signals using a FPGA. The generator displays successfully the intended pattern sequence(REDlongrightarrowGREENlongrightarrowBLUElongrightarrowWHITElongrightarrowBLACK) and the result are demonstrated at the conference site. The advantage of the use of the proposed generator is the simplification of production equipments and the pattern generator.

In this paper, we try to make a systematic analysis for the vertical crosstalk on LCD driving. The vertical Crosstalk is one of the inevitable phenomena in the high-resolution TFT-LCDs. At first, the vortical Crosstalk is defined. Considering the structure of the pixels and data lines, a LCD (a pixel) is modeled electrically by its' equivalent circuit. The circuit is verified by experiment and used to find the cause of the vertical Crosstalk.

In this paper, a new automated defects detection method for a TFT-LCD panel is presented. An input image is preprocessed to lessen small abnormal noises and non-uniformity of the image. The adaptive multi-thresholds are used to detect Muras, which are the major defects occurred on TFT-LCD panels. Those are determined adaptively depending on the brightness and the brightness distribution of a local block. For the synthetic images and real Mura images, the proposed algorithm can effectively detect Muras in a reasonable time.

To inspect point-defect in LCD pannel, calculate period and eliminated pattern. And then find point-defect to compare block image with each period. First processing, Founded over point defects. To reduce wrong point defect. Next, label point-defects and eliminated not surpass fixed limit-size.

Basically, the mis-convergence, which is inevitable in CRT Projection TV. is the degree of deviation of red and blue from green beam. The cause of mis-convergence is the change of magnetic field and electrical characteristic in deflection circuits and convergence amplification circuit. A new and easily implementable mis-convergence compensation algorithm is presented in this paper. The proposed algorithm does not needs any compensation devices. It uses only TV OSD and a remote controller and anyone who wants to compensate can easily correct the mismatch. Through real compensation experiments, it is found that the proposed algorithm is useful and effective one.

In PDP(Plasma Display Panel), sub-field method is used to implement gray scale. Each sub-field has different periods. And Every gray level has information of which sub-field has to be displayed. This is called sub-field mapping. There are several sub-field mapping values in some gray levels. So, it is possible to select best choice in this paper, we propose new sub field mapping method using a fuzzy inference system to select best sub-field mapping values in accordance with input image and environment temperature. In order to implement fuzzy system, we used MATLAB fuzzy inference editor.

We propose the system that automatically adjusts the white balance on display products to a standard value according to several nations. We replace manual or semi-auto method with fully automatic method using windows application program. And we use RS-232C serial interface to communicate PC with display products which we want to adjust white balance. The PC generates patterns for measuring color information and Color Analyzer measures color and brightness. This value is transmitted through RS-232C serial interface to PC. The PC's algorithm analyzes this information and then decides which RGB Gain value is best for optimal white balance. This RGB Gain value is transmitted through RS-232C serial interface to display products. The modified color value is measured again and feed back to PC. This sequence is repeated until optimum white balance is obtained.

In this paper, we propose impr oved binary tree vector quantization based on spatial sensitivity which is one of the human visual properties. We combine the weights based on spatial masking effect according to changes of three primary colors in blocks of images with the process of splitting nodes using eigenvector in binary tree vector quantization. The test results show that the proposed method generates the quantized images with fine color and performs better than the conventional method in terms of clustering the similar regions. Also the proposed method can get the better result in subjective qualify test and PSNR.

This paper presents a image modification algorithm using the convergence camera model and the two perspective projection matrixes of the original cameras. Any pair of images can be transformed se that epipolar lines are parallel and horizontal in each image. The advantage of modification is that a 2-D search problem is reduced to a 1-D search problem. Reconstruction can be performed directly from the rectified images.

A new color image segmentation scheme is presented in this paper. The proposed algorithm consists of image simplification, region labeling and color clustering. The vector-valued diffusion process is performed in the perceptually uniform LUV color space. We present a discrete 3-D diffusion model for easy implementation. The statistical characteristics of each labeled region are employed to estimate the number of total clusters and agglomerative hierarchical clustering is performed with the estimated number of clusters. Since the proposed clustering algorithm counts each region as a unit, it does not generate oversegmentation along region boundaries.

Among digital halftoning methods, error diffusion is a procedure for generating high quality bilevel images from continuous-tone images but blurs the edge information in the bilevel images. To solve the problem, we propose the edge enhanced error diffusion using the edge information of the original images. The edge enchanted weights is computed by adding local characteristic weights and input pixels multiplied a constant. Also, we combined the edge enhanced method with the adaptive error diffusion using human spatial and frequency perception characteristic. The performance of the proposed method is compared with conventional method by measuring the edge correlation. The halftoned images applied the proposed method get more fine quality due to the enchanced edge and better quality in halftoned image. And the detailed edge is preserved in the halftoned images by the proposed method.

In this paper, we propose a new saturation enhancement algorithm which is processed on the new color space, called Normalized YCbCr(NYCbCr). The algorithm consists of two processing unit. One is color space conversion from YCbCr to NYCbCr, and the other is using adaptive saturation mapping function(ASMF). NYCbCr color space is designed to prevent shortcomings such as luminance and hue shift of YCbCr color space and by saturation enhancement. ASMF is effective to enhance saturation properly for each image and to protect low saturation regions of color images from over-saturation. we verified our method using several color images. Experimental results show that the proposed method enhance the saturation with minimizing Luminance and Hue shift.

원격감지(remote sensing) 기술의 비약적인 발전과 함께 다중분광 영상데이터의 분광대역수가 급속히 증가하고 있다. 대역수의 증가로 영상데이터의 양이 급격히 증가하게 되고, 이에 따라 이들 데이터를 처리하기 위해서는 처리속도가 빠른 영상 처리 기술이 필요하게 되었다. 분광 대역수를 줄여 빠르게 처리하는 한가지 방법으로 널리 사용되고 있는 것이 주성분변환이다. 본 논문에서는 주성분변환에 대한 처리방법에 대해 논한 후, 다중분광 영상데이터를 주성분 변환한 주성분 영상데이터를 분석하였다. 또한 주성분 영상데이터를 최대유사법으로 분류하고 그 결과를 분석하였다.

This paper proposes a method to speed up the document image binarization using a water flow model. The proposed method extracts the region of interest (ROI) around characters from a document image and restricts pouring water onto a 3-dimensional terrain surface of an image only within the ROI. The amount of water to be filled into a local valley is determined automatically depending on its depth and slope. Then, the proposed method accumulates weighted water not only on the locally lowest position but also on its neighbors. Finally, the depth of each pond is adaptively thresholded for robust character segmentation. Experimental results on real document images shows that the proposed method has attained good binarization performance as well as remarkably reduced processing time compared with that of the existing method based on a water flow model.

본 연구에서는 KM-Net(Hidden Markov Network)을 다양한 태스크에의 적용과 화자의 특성을 효과적으로 나타내기 위해 HM-Net 음성인식 시스템에 MLLR(Maximum Likelihood Linear Regression) 적응방법을 도입하였으며, HM-Net 학습 알고리즘을 개량하여 회귀클래스 생성방법을 제안한다. 제안방법은 PDT-SSS(Phonetic Decision Tree-based Successive State Splitting) 알고리즘의 문맥방향 상태분할에 의한 상태레벨 공유를 이용한 방법으로 새로운 화자로부터 문맥정보와 적응화 데이터의 발성 양에 의존하여 결정된 많은 적응 파라미터들을(평균, 분산) 자유롭게 제어할 수 있게 된다. 제안방법의 유효성을 확인하기 위해 국어공학센터(KLE) 452 음성 데이터와 항공편 예약관련 연속음성을 대상으로 인식실험을 수행한 결과, 전체적으로 음소인식의 경우 평균 34-37％, 단어인식의 경우 평균 9％, 연속음성인식의 경우 평균 7-8％의 인식성능 향상을 각각 보였다. 또한 적응화 데이터의 양에 따른 인식성능 비교에서, 제안방법을 적용한 인식 시스템이 적응 데이터의 양이 적은 경우에도 향상된 인식률을 보였으며. 잡음을 부가한 음성에 대한 적응화 실험에서도 향상된 인식성능을 보여 MLLR 적응방법의 특성을 만족하였다. 따라서 MLLR 적응방법을 도입한 HM-Net 음성인식 시스템에 제안한 회귀클래스 생성방법이 유효함을 확인한 수 있었다.

이 논문에서는 이산형 상태공간 모델에 대한 $H_2$, $H_{\infty}$, 및 혼합 $H_{\infty}$ FIR 필터를 선형행렬부등식(LMI)를 이용하여 제안한다. 제안되는 필터는 FIR 구조로서 $H_2$ 및 $H_{\infty}$ 관점에서의 성능기준을 만족함과 더불어 선형성 및 불편향성의 특성을 지니고, 초기 상태에 관한 정보를 필요로 하지 않는다. 그리고 FIR 구조로 인해 기존의 FIR 형태의 필터에 비해 불확실성에 대해 보다 견실하며 빠른 수렴성을 갖는다. 모의 실험을 통해 이러한 장점을 예시한다.

This paper proposes a new algorithm for cloudy area detection using K-Means and GHA (Generalized Hebbian Algorithm). K-Means is one of simple classification algorithm, and GHA is unsupervised neural network for data compression and pattern classification. Proposed algorithm is based on block based image processing that size is l6$\times$l6. Experimental results shows good performance of cloudy area detection except blur cloudy areas.

In this paper, the encoder architecture of 3-D wavelet transform based on lifting scheme is designed. Architecture, here, 3 level wavelet transform for spatial decomposition and 2 level wavelet transform for temporal decomposition is adopted with efficient computation.

This paper proposes an adaptive background image generation method based on the frame difference for traffic monitoring. The performance of the conventional method is limited when there are more vehicles due to traffic Jam. To improve on this, we use frame differencing to separate vehicles from background in frame differencing, we adopt selective approach by using part of the image not considered as vehicle fer extraction of background. The proposed method generates background more efficiently than conventional methods even in the presence of heavy traffic.

Most Embedded Web Camera Server products currently deployed on the market adopt JPEG for compression of video data continuously acquired from the cameras. However, JPEG does not efficiently compress the continuous video stream, and is not appropriate for the Internet where the transmission bandwidth is not guaranteed. In our previous work, we presented the result of designing and implementing an embedded web camera streaming server using MPEG4 codec. But the server in our previous work did not show good performance since one CPU had to both compress and process the network transmission. In this paper, we present our efforts to improve our previous result by using dual CPUs, where DSP is employed for data compression and StrongARM is used for network processing. Better performance has been observed, but it is found that still more time is needed to optimize the performance.

The KOMPSAT-II has a MSC(Multi-Spectral Camera) payload for earth observatory. The image data acquired during the pass over the Korean Peninsula can be sent to the ground station directly. But the image data out of the contact range should be stored temporally for later transmission. The KOMPSAT-II has a device for this purpose called the DCSU(Data Compression and Storage Unit) and the DCSU also performs compression functions for saving storage space and transmission time to send image data to the ground station. In this paper, we'd like to introduce the DCSU memory structures and operation.

Recently, the IDS(Intrusion Detection System) using a video camera is an important part of the home security systems which start gaining popularity. However, the video intruder detection has not been widely used in the home surveillance systems due to its unreliable performance in the environment with abrupt illumination change. In this paper, we propose an effective moving edge extraction algorithm from a sequence image. The proposed algorithm extracts edge segments from current image and eliminates the background edge segments by matching them with reference edge list, which is updated at every frame, to find the moving edge segments. The test results show that it can detect the contour of moving object in the noisy environment with abrupt illumination change.

영상압축 방법인 ITU-T H.263 부호화 기법에 비해 H.263+에서의 주요 특징 중 하나는 에러 강인성으로, 그 중에서 계층부호화 모프는 한 개의 송신 영상에 대해서도 복호기와 전송로에 따라 여러 가지 화질의 재생 영상을 얻을 수 있는 기능, 즉 에러 및 패킷 손실 발생 가능 채널에서 비디오 정보 진송시 디코더측에서 사용 가능한 다양한 비트율, 해상도 디스플레이율을 허용함으로써 비디오 정보에 대한 전송을 향상시키는 기법이다. 계층 부호화 에러은닉기법은 기본계층에 대해서는 QoS 보장에 의해 무손실 전송이 가능하므로 모든 프레임에 에러가 발생할 수 있는 단일계층 에러은닉보다 고급계층에서 이용한 수 있는 정보가 더 많다는 점에서 더 유리한 이점을 가지고 있다. 본 논문에서는 SNR 계층부호화된 고급계층에 대해서 이용 가능한 데이터 손실이 없는 기본계승과 고급계층에 상관관계를 분석하여 계층형 부호화에 맞는 에러은닉기법을 제시한다.

In this paper, the efficient motion estimation and compensation method for 3 dimensional wavelet transform is proposed. Recently, since the compression performance and scalable functionality are provided by wavelet transform, many researches have been carried out for applying to the video compression. For the temporal filtering, motion estimation and compensation techniques are used, but the unconnected pixels, which are produced by motion compensation result into the degradation of coding performance and quality of the picture. For the efficient motion compensated temporal filtering by reducing the number of these unconnected pixels, we propose the variable block size motion estimation and compensation method. Also we propose a method that determines the block size using rate-distortion optimization technique according to the local characteristics of the frame. The simulation results show the improved performances than the MPEG-4 scalable coding methods and the 3 dimensional wavelet coding methods using fixed block size motion estimation and compensation.

When we search database with a query image, the retrieval efficiency will vary from each kind of descriptor. Even the best representative descriptor, it results a few useless images that don't match with query image. This type of error can be reduced by adopting another descriptor which extracts features in different way. At present, the choice of descriptors is base on intuitive and experimental method. By theoretic accessing to the problem of descriptor choice, we can solve the given problem in the objective and rational way. In this study, we intend to make a composite of descriptors that can reduce retrieval error by adopting principal component analysis.

본 논문은 동영상 압축 표준 H.264에서 화면간 예측 부호화시 부호화 효율을 높일 수 있는 참조 화면 선택에 대한 기법을 제안하였다. 선택에 대한 문제를 참조 화면 간 움직임 정보를 고려해서 가장 유사한 두 참조 화면을 구하는 것으로 해결하고자 하였고 움직임 정보에 대한 고려는 부호화시 계산되는 R-D 비용함수 값을 사용함으로서 부가적인 계산을 최소화 하고자 하였다. 실험결과 기존의 슬라이딩 윈도우 방식에 비해 Bit Rate 감소로 인한 부호화 효율개선을 얻을 수 있었다.

In this paper, we present a hierarchical approach of an enhanced active shape model for video tracking. Kalman filter is used. To estimate a dynamic shape in video object tracking. The experimental results show that the proposed hierarchical active shape model using Kalman filter is efficient.

This paper propose a extracting method of the region for image using segmentation and edge information. First propose algorithm extract information using canny edge detector and the image was divided by watershed segmentation. And it extract the mage with edge information by merging region. Finally we compare the proposed method with levelset method. In the result proposed method not only extract the image with accurate region but also reduce operation time.

Compared with other features of the image, color features are less sensitive to noise and background complication. Besides, this adding to object segmentation has more accuracy of image retrieval. This paper presents object segmentation and HAQ(Histogram Analysis and Quantization) algorithm approach to extract features(the object information and the characteristic colors) of an image. The empirical results shows that this method presents exactly spatial and color information of an image as image retrieval's feature.

DC억압능력이 없거나 부족한 코드에 만족할 만한 DC억압능력을 갖도록 하는 방법은 DC 제어 비트의 사용, Dual Code의 사용, Multimode Code의 사용 등이 있다. 어떤 방법이든 부가 비트가 사용되지만, 그 중에서 멀티모드 변조코드는 우수한 DC억압능력과 높은 코드효율을 갖고 있음에도 불구하고 복잡한 하드웨어와 높은 에러 전파율을 갖는 단점도 있다. 본 논문에서 제시하는 멀티모드 변조코드의 특징은 데이터열의 다중화를 위해 의사 스크램블 기법을, 다중화된 데이터열의 변조를 위해서는 DC-free RLL 변조코드를 사용한다. 의사 스크램블에 의한 데이터열의 다중화는 데이터를 복조할 때 에러전파 확률을 떨어뜨리는 효과가 있고 다중화된 데이터열의 변조를 위한 DC-free RLL 변조코드의 사용은 DC억압능력을 향상시키고 하드웨어가 훨씬 간단해진다.

This paper proposes the multi-level vector error diffusion for smear artifact reduction in the boundary regions. Smear artifact mainly results from a large accumulation of quantization error. Accordingly, to reduce these artifacts, the proposed method excludes the large quantization error in the error diffusion process by comparing the magnitude of the error vector with predetermined first threshold. In addition, if the vector norm of the difference between the error adjusted input vector and the primary co]or that has minimum vector norm for the error adjusted input vector is larger than second threshold, the error is excluded. As a result, the proposed method reduce smear artifact in the boundary region and produces visually pleasing halftone pattern.

본 논문에서는 광 저장장치의 기록밀도가 증가할 경우 발생하는 여러 가지 문제점들에 대응하는 PRML 시스템 구현 방법을 제안하고 기존의 시스템의 성능과 제안된 시스템의 성능을 비교한다. 기존의 채널 적응기법 및 비터비 복호화기에 대비해 보다 종은 성능을 가지는 구조의 비선형 등화방식과 비터비 예상 레벨 적응 알고리즘을 이용한 새로운 구조를 제안하고, 23Gbyte 저장용량을 가지는 Blu-ray 디스크에 28GB의 데이터를 기록하여 기록 기록밀도가 증가한 실제 디스크에 대한 실험 결과에 대하여 논한다.

Skew is inevitably occurred in a scanned document image Thus, character recognition systems are generally very sensitive to a skew angle. In this paper, we propose a robust slant correction algorithm based on dithering and estimating vortical transition. Character strings are segmented by projecting the vertical transition point and the slant corrected image. The segmentation method using the vertical transition point can effectively split the character strings touching vertically each other. Experimental results show that the proposed method has achieved robust slant correction and good performance of character string segmentation.

We suggest 2 dimensional Fast Fourier Transform using Polynomial Transform and integer Fast Fourier Transform. Unlike conventional 2D-FFT using the direct quantization of twiddle factor, the suggested 2D-FFT adopts implemented by the lifting so that the suggested 2D-FFT is power adaptable and reversible. Since the suggested FFT performg integer-to-integer mapping, the transform can be implemented by only bit shifts and auditions without multiplications. In addition. polynomial transform severely reduces the multiplications of 2D-FFT. While preserving the reversibility, complexity of this algorithm is shown to be much lower than that of any other algorithms in terms of the numbers of additions and shifts.

This paper describes a novel method for image contrast enhancement by controlling gamma curve in AMLCD. The key idea is to automatically manipulate gamma voltage in accordance with the image data distribution. This method is applied to 17" SXGA LCD monitor module. The contrast ratio and the brightness are enhanced respectively by about 3 times and 1.7 times, by using the proposed method.

In this paper, we propose a method that acquires proper image for barcode reading on camera-based barcode reading system. In the camera system, there is serious blurring problem, and to get proper output of the barcode, it must be solved. To solve it, we proposed two methods. The first one is to make the camera movable, so it can change the distance of focus. The other is software approach, and threshold value is revised through regression analysis. Using such process, we can notice that blurring problem can be overcome.

A conventional correlation target tracker is analysed with a simple mathematical approach. And, we will propose a correlation measure with selective attentional property in order to overcome the false-peak problem of the conventional methods. Various experimental results show that the proposed correlation measure is able to reduce considerably the probability of false-peaks degraded by the correlation between background images of a reference block and a distorted and noisy sensor input image.

In this paper, we propose an iterative mixed norm image restoration algorithm using multi regularization parameters. A functional which combines the regularized l$_2$ norm functional and the regularized l$_4$ functional is proposed. The smoothness of each functional is determined by the regularization parameters. Also, a regularization parameter is used to determine the relative importance between the regularized l$_2$ functional and the regularized l$_4$ functional. An iterative algorithm is utilized for obtaining a solution and its convergence is analyzed.

Recently, Ryu et al. proposed a multiple target DOA tracking algorithm, which has good features that it has no data association problem and simple structure. But its performance is seriously degraded in the low signal-to-noise ratio. In this paper, a measurement fusion method is presented based on ML(Maximum Likelihood), and the new DOA tracking algorithm is proposed by incorporating the presented fusion method into Ryu's algorithm. The proposed algorithm has a better tracking performance than that of Ryu's algorithm, and it sustains the good features of Ryu's algorithm.

이 논문에서는 CIC(Cascaded Integrator-Comb) 데시메이션 필터의 주파수응답을 향상시키는 새로운 필터 구조를 제안한다. 기존의 방식은 CIC 필터의 통과대역 특성은 향상시키나, Aliasing 대역의 감쇠특성은 조금 나빠지는 단점을 갖고 있다. 그러나 이 논문에서는 통과대역의 특성은 기존의 방식보다 우수하며, 동시에 Aliasing 대역의 감쇠특성도 매우 향상시키는 4차의 보간필터를 제안한다. 제안된 필터는 4차의 필터이나 곱셈이 1개만 필요한 구조이므로 부가적인 연산량이 적으며, 또한 선형위상의 특성을 갖고 있으므로 CIC 필터의 선형위상 특성을 그대로 유지한 수 있다. 기존의 주파수응답 향상 기법들과 특성개선 효과를 비교하였다.

This paper presents a new fractal coding scheme to find more optimal transformation by estimation of the optimal attractor. The conventional fractal coding schemes based on the collage theorem obtain the transformation to minimize the distance between an original image and its collage image. Heavy computation is why the schemes widely adopt the theorem. In other words, the optimal transformation can be obtained after the attractors of all the possible transformations are generated and then compared with an original image. It is clear that this process is not practical. Therefore, we introduce a sub-optimal scheme that provides better transformation than the conventional scheme, relieving the complexity problem in the optimal transformation. In a simple case, the optimal transformation can be obtained considering all the attractors and then our scheme is compared with the optimal. In general cases not to be able to find the optimal, our scheme is also evaluated and compared with the conventional schemes.

The current paper proposes a new vertex selection scheme for polygon-based contour ceding. To efficiently characterize the shape of an object, we incorporate the curvature information in addition to the conventional maximum distance criterion in vertex selection process. The proposed method consists of “two-step procedure.” At first, contour pixels of high curvature value are selected as key vertices based on the curvature scale space (CSS), thereby dividing an overall contour into several contour-segments. Each segment is considered as an open contour whose end points are two consecutive key vertices and is processed independently. In the second step, vertices for each contour segment are selected using progressive vertex selection (PVS) method in order to obtain minimum number of vertices under the given maximum distance criterion ( $D_{MAX}$). Experimental results are presented to compare the approximation performances of the proposed and conventional methods.s.

On-going convergence between broadcasting and communication is currently bringing in lots of drastic changes to multimedia services to the home. The FTTH-based home networks will be the interface among the digital home appliances for broadcast digital contents as weil as for broadband data connection. In this paper, we discuss several design issues for home gateway architecture that effectively distributes realtime broadcasting contents (from either the headend in the access network or the satellite/terrestrial broadcasts) to heterogeneous devices in the home network.

Current paper introduces the additional multimedia module for digital TV. The module is developed for displaying the image captured by digital still camera, camcorder, or PC in the digital TV. For these purpose, the module has the interface circuit for accessing five media type of memory cards. It decodes JPEG, BMP, or TIFF image data saved in the memory card and converts the image data to analog RGB signal. It also supports three types of output image size from HD to WXGA resolution. So the introduced module could be adopted In the most of digital TV.

Commercial video scope use CCD sensor and frame grabber for image capture and A/D interface but application limited by input resolution and high cost. In this paper we introduce portable video scope using CMOS sensor, USB pen and tuner card (low frame grabber) in place of commercial CCD sensor and frame grabber. Our video scope serves as an essential link between advancing commercial technology and research, providing cost effective solutions for educational, engineering and medical applications across an entire spectrum of needs. The software implementation is done using Direct Show in second version after initial trials using First version VFW (video for window), which gave very low frame rate. Our video scope operates on windows 98, ME, XP, 2000. The drawback of our video scope is crossover problem in output images caused due to interpolation, which has to be rectified for more efficient performance.

In this paper, we propose a face region extraction algorithm using color information and projection. After the extraction of face candidate image using adaptive color information, we project it into vertical direction to estimate the width of the face. Then the redundant parts of the face are efficiently removed by using the estimated face width. And the width information of the face is used at the horizontal projection step to extract the height of the face, and non-face region such as the neck and some background regions, which are represented as the similar skin color, effectively eliminated. From the experiment results for the various images, the proposed algorithm shows more accurate results than the conventional algorithm.

In this paper, we present a novel idea to integrate low cost Inertial Measurement Unit(IMU) and Differential Global Positioning System (DGPS) for Telematics applications. As well known, low cost IMU produces large positioning and attitude errors in very short time due to the poor quality of inertial sensor assembly. To conquer the limitation, we present a bimodal approach for integrating IMU and DGPS, taking advantage of positioning and orientation data calculated from CCD images based on photogrammetry and stereo-vision techniques. The positioning and orientation data from the photogrammetric approach are fed back into the Kalman filter to reduce and compensate IMU errors and improve the performance. Experimental results are presented to show the robustness of the proposed method that can provide accurate position and attitude information for extended period for non-aided GPS information.

In this paper, we describe new method to detect face in real-time. We use color information, edge information, and binary information to detect candidate regions of eyes from input image, and then extract face region using the detected eye pall. We verify both eye candidate regions and face region using Support Vector Machines(SVM). It is possible to perform fast and reliable face detection because we can protect false detection through these verification processes. From the experimental results, we confirmed the proposed algorithm shows very excellent face detection performance.

In this paper we propose AODC(Adaptive Over Driving Circuit) to reduce motion blurring and compensate color accuracy on LCD monitor. The proposed method has been experimented on adaptive ODC by motion grade after detection of moving objects. Also, we discriminated motion grade by mean value's degree of differential value between current frame and previous frame to detect moving objects. It has been found that the proposed method provides better performance than the conventional one from the point of view on the blurring of movement objects through the experiment. Its results show that the proposed method is very suitable for a motion blurring reduction on LCD monitor.

In this paper, we evaluated error correction performance and recording density of an optical disc system. The performance of Low-Density Parity Check code (LDPC) is compared to the HD-DVD (BD) ECC. The recording density of optical disc can be increased by reducing the redundancy of the user data. Moreover, since the correction capability of LDPC with decreased redundancy is better than that of BD, the recording density can also be increased by reducing the mark length of the data on the disc surface.

This paper proposes a new color feature and a corresponding distance measure for content-based retrieval of design images such as trade marks, pattens, logos, textile images, and icons. Simulation results with textile images show that the proposed method outperforms the traditional color-based retrieval methods which was originally proposed fer content-based retrieval of natural images.

MIMO-OFDM 시스템은 다중 안테나 송신을 통하여 시스템의 capacity를 극대화 하지만, 이를 위해서는 정확한 채널계수 값의 추정을 필요로 한다. 본 논문에서는 MIMO-OFDM시스템을 위한 채널계수 추정기법으로서 Space-Time Coding(STC)에 기반을 둔 방식을 제안한다. 제안된 채널 계수 추정 기법을 위해 필요한 심볼 구조, frame 구조를 IEEE 802.11a 시스템에 적용하였다. VBLAST 기법이 적용된 IEEE 802.11a 기반의 4 by 4 MIMO-OFDM 시스템에 대한 모의 실험을 통하여 제안된 방법의 성능과 기존의 방법의 성능을 coded BER를 통해 비교 분석해 보았다.

The electrical current generated by heart creates not only electric potential but also a magnetic field. We have observed electrophysiological phenomena of the heart by measuring tangential components of magnetocardiogram(MCG) using 61 channel superconducting quantum interference device(SQUD) system. In this paper, we developed a new analysis method, which is based on the theory of electromagnetic field. We show some differences of the current direction between the normal MCG and the abnormal(ischemic heart disease) MCG.

In the field of content-based image retrieval, various mathematical low-level features have been proposed to describe the perceptual content of images. Since most of the features are assumed to be independent of each other, one feature is extracted from images without any consideration of the other features. Recently proposed CCE and SCFT taking advantage of the correlation between color and texture have shown relatively good performance. In this paper, the performance of CCE, SCFT, and the traditional regular weighted comparison method are evaluated. Simulation results with natural images have shown that CCE outperforms the other methods.

Voice is one of the promising biometrics because it is one of the most convenient ways human would distinguish someone from others. The target of speaker verification is to divide the client from imposters. Support Vector Machine(SVM) is in the limelight as a binary classifier, so it can work well in speaker verification. In this paper, we combined SVM with genetic algorithm(GA) to reduce the dimensionality of input feature. Experiments were conducted with Korean connected digit database using different feature dimensions. The verification accuracy of SVM with GA is slightly lower than that of SVM, but the proposed algorithm has greater strength in the memory limited systems.