A chromatic contour detector based on abrupt change techniques

A method for segmenting colored images based on contour detection is described. It involves the use of techniques based on abrupt change detection used in signal processing and control systems. This work is based on a theoretical framework: differential geometry, tensors, Neyman-Pearson's optimal decision. Good results are obtained with Gaussian and exponential noises. This method also provides a very good location of contours.

[1]  Rachid Deriche,et al.  An efficient method to build early image description , 1988, [1988 Proceedings] 9th International Conference on Pattern Recognition.

[2]  King-Sun Fu,et al.  A survey on image segmentation , 1981, Pattern Recognit..

[3]  Y. Ohta.  Knowledge-based interpretation of outdoor natural color scenes , 1998 .

[4]  Larry S. Davis,et al.  A survey of edge detection techniques , 1975 .

[5]  Theodosios Pavlidis,et al.  Structural pattern recognition , 1977 .

[6]  Silvano Di Zenzo,et al.  A note on the gradient of a multi-image , 1986, Comput. Vis. Graph. Image Process..

[7]  Henri Maître,et al.  A robust method for picture segmentation based on a split-and-merge procedure , 1986, Comput. Vis. Graph. Image Process..

[8]  Michèle Basseville,et al.  Detection of abrupt changes , 1993 .

[9]  Keith Phillips,et al.  Applications of Vector Fields to Image Processing , 1983, IEEE Transactions on Pattern Analysis and Machine Intelligence.