DOI QR코드

DOI QR Code

장면 텍스트 추출을 위한 캐니 연산자의 적응적 임계값을 이용한 AEMSER

AEMSER Using Adaptive Threshold Of Canny Operator To Extract Scene Text

  • Park, Sunhwa (Gyeongsang Univ. Dept. of Computer Science and Graduate School of CCBM) ;
  • Kim, Donghyun (Gyeongsang Univ. Dept. of Computer Science) ;
  • Im, Hyunsoo (Gyeongsang Univ. Dept. of Computer Science) ;
  • Kim, Honghoon (Gyeongsang Univ. Dept. of Computer Science) ;
  • Paek, Jaegyung (Gyeongsang Univ. Dept. of Computer Science) ;
  • Park, Jaeheung (Gyeongsang Univ. Dept. of Computer Science and Graduate School of CCBM) ;
  • Seo, Yeong Geon (Gyeongsang Univ. Dept. of Computer Science and Graduate School of CCBM)
  • 투고 : 2015.11.18
  • 심사 : 2015.12.31
  • 발행 : 2015.12.31

초록

장면 텍스트 추출은 현대 스마트 시대에서 쏟아져 나오는 다양한 영상 기반 응용에 중요한 정보를 제공하기 때문에 중요하다. 기본적인 MSER(Maximally Stable Extremal Regions) 추출 후에 캐니 연산자를 이용하여 경계를 강화시키는 Edge-Enhanced MSER은 텍스트 추출 측면에서 뛰어난 성능을 보인다. 하지만 캐니 연산자의 임계값 설정에 따라 Edge-Enhanced MSER의 결과영상이 다르게 나타나므로 임계값 설정을 계산하는 방법이 필요하다. 본 논문에서는 캐니 연산자의 임계값을 설정하는 방법 중 히스토그램의 중앙값을 이용하여 경계를 추출하고 이를 Edge-Enhanced MSER에 적용한 AEMSER(Adaptive Edge-enhanced MSER)을 제안한다. 이 방법은 명확한 경계에 대해서만 영역을 추출하기 때문에 기존의 방법보다 더 좋은 결과영상을 얻을 수 있다.

Scene text extraction is important because it offers some important information on different image based applications pouring in current smart generation. Edge-Enhanced MSER(Maximally Stable Extremal Regions) which enhances the boundaries using the canny operator after extracting the basic MSER shows excellent performance in terms of text extraction. But according to setting the threshold of the canny operator, the result images using Edge-Enhanced MSER are different, so there needs a method figuring out the threshold. In this paper, we propose a AEMSER(Adaptive Edge-enhanced MSER) that applies the method extracting the boundary using the middle value of histogram to Edge-Enhanced MSER to get the canny operator's threshold. The proposed method can acquire better result images than the existing methods because it extracts the area only for the obvious boundaries.

키워드

참고문헌

  1. Li, Yao, et al., "Characterness: an indicator of text in the wild.", Image Processing, IEEE Transactions on 23.4 : pp.1666-1677, 2014. https://doi.org/10.1109/TIP.2014.2302896
  2. Jung, Keechul, et al., "Text information extraction in images and video: a survey.", Pattern recognition 37.5, pp.977-997, 2004. https://doi.org/10.1016/j.patcog.2003.10.012
  3. Kang, Le, et al., "Orientation robust text line detection in natural images.", Computer Vision and Pattern Recognition (CVPR), IEEE, 2014.
  4. Sung, Myung-Chul, et al., "Scene Text Detection with Robust Character Candidate Extraction Operator.", 13th ICDAR, 2015.
  5. Canny, John, "A computational approach to edge detection." IEEE Trans. Pattern Anal. Mach. Intell., vol. 8, pp.679-698, 1986.
  6. Fang, Mei, et al., "The study on an application of otsu Operator in canny operator." International Symposium on Information Processing (ISIP). 2009.
  7. Chen, H., et al., "Robust text detection in natural images with edge-enhanced maximally stable extremal regions." In Image Processing (ICIP), 18th IEEE International Conference on, pp.2609-2612, Sep. 2011.
  8. Upadhyay, Nishchal Gyan, and Kamlesh Lakhwani, "Edge Detection Using Fuzzy Approach Involving Automatic Threshold Generation.", International Journal Of Scientific & Techonology Research Vol. 2, Iss. 7, pp.128-131, July 2013.
  9. Epshtein, Boris, et al., "Detecting text in natural scenes with stroke width transform.", Computer Vision and Pattern Recognition, IEEE, 2010.
  10. Kerry D. Wong, "Canny Edge Detection Auto Thres holding", http://www.kerrywong.com/2009/05/07/canny-edge-detection-auto-thresholding/, Sep. 2015

피인용 문헌

  1. Performance comparison of Image De-nosing Techniques based on Color Model Transformation vol.18, pp.8, 2015, https://doi.org/10.9728/dcs.2017.18.8.1641