• Title/Summary/Keyword: 오디오 표식

Search Result 3, Processing Time 0.032 seconds

Study on data augmentation methods for deep neural network-based audio tagging (Deep neural network 기반 오디오 표식을 위한 데이터 증강 방법 연구)

  • Kim, Bum-Jun;Moon, Hyeongi;Park, Sung-Wook;Park, Young cheol
    • The Journal of the Acoustical Society of Korea
    • /
    • v.37 no.6
    • /
    • pp.475-482
    • /
    • 2018
  • In this paper, we present a study on data augmentation methods for DNN (Deep Neural Network)-based audio tagging. In this system, an audio signal is converted into a mel-spectrogram and used as an input to the DNN for audio tagging. To cope with the problem associated with a small number of training data, we augment the training samples using time stretching, pitch shifting, dynamic range compression, and block mixing. In this paper, we derive optimal parameters and combinations for the augmentation methods through audio tagging simulations.

A Watermarking Scheme to Extract the Seal Image without the Original Image (원본정보 없이 씰영상의 추출이 가능한 이미지 워터마킹 기법)

  • Kim, Won-Gyum;Lee, Jong-Chan;Lee, Won-Don
    • The Transactions of the Korea Information Processing Society
    • /
    • v.7 no.12
    • /
    • pp.3885-3895
    • /
    • 2000
  • The emergence of digital imaging and digital networks has made duplication of original artwork easier. In order to protect these creations, new methods for signing and copyrighting visual data are needed. In the last few years, a large number of schemes have heen proposed for hiding copyright marks and other information in digital image, video, audio and other multimedia objects. In this paper, we propose a technique for embedding the watermark of visually recognizable patterns into the frequency domain of images. The embedded watermark can be retrieved from the decoded sequence witbout knowledge of the original. Because the source image is not required to extract the watermark, one cannot make the fake original that is invertible to watermarking scheme from the waternlarked image. In order to recover the embedded signature data without knowledge of the original, a prediction of the original value of the pixel containing the information is needed. The prediction is based on a averaging of amplitude values in a neighborhood around the pixel itself. Additionally the projxJsed technique could survive several kinds of image processings including JPEG lossy compression.

  • PDF