A Fast Hybrid Color Segmentation Method

We introduce a new, very general method to achieve stable and very fast color segmentation. This method works on diierent hierarchical data structures and combines local bottom-up region growing segmentation with top-down separation techniques. This method has been applied to implement an object detector. As our method supports exploitation of inherent parallelism, a real-time object detection has been designed and is in the implementation phase.

[1]  T. Kanade,et al.  Color information for region segmentation , 1980 .

[2]  Kannan,et al.  ON IMAGE SEGMENTATION TECHNIQUES , 2022 .

[3]  Mehmet Celenk,et al.  A color clustering technique for image segmentation , 1990, Comput. Vis. Graph. Image Process..

[4]  Lutz Priese,et al.  An asynchronous parallel algorithm for picture recognition based on a transputer net , 1990 .

[5]  John E. Howland,et al.  Computer graphics , 1990, IEEE Potentials.

[6]  Keith Price,et al.  Picture Segmentation Using a Recursive Region Splitting Method , 1998 .

[7]  Matti Pietikäinen,et al.  Color segmentation by hierarchical connected components analysis with image enhancement by symmetric neighborhood filters , 1990, [1990] Proceedings. 10th International Conference on Pattern Recognition.

[8]  R. Stephenson.  A and V , 1962, The British journal of ophthalmology.

[9]  Ron Gershon,et al.  Aspects of perception and computation in color vision , 1985, Comput. Vis. Graph. Image Process..

[10]  Azeddine Beghdadi,et al.  Contrast enhancement technique based on local detection of edges , 1989, Comput. Vis. Graph. Image Process..