Extracting the Contours of a Zebra Against a Complex Striped Background Coping with Portions of Subjective Contours

H. Muramatsu, H. Banno, T. Tanaka, and N. Sugie (Japan)


Intelligent Information System, AI Algorithm, Computer Vision, Signal and Image Processing, Contour Extraction, Subjective Contour


When a vision system tries to recognize objects in an image, the extraction of the contours constitutes a basic and essential processing. However, it is difficult to extract objects whose contours include subjective contours by conventional processing alone. Yasuda et al. proposed a basic algorithm to deal with typical cases of subjective contours. Afterwards, the algorithm was improved to cope with more complex figures. In the case of very complex figures, even the algorithm could not generate some of correct contours that humans perceive easily. So we extend further the algorithm to extract most of the correct subjective contours.

Important Links:

Go Back