Using Local Features in a Neural Network Based Gray-Level Reduction Technique

This paper proposes a new method for reduction of the number of gray-levels in an image. The proposed approach achieves gray-level reduction using the image gray-levels and additional local spatial features. Both gray-level and local feature values feed a self-organized neural network classifier. The final image has not only the dominant image gray-levels, but also has texture approaching the image local characteristics used. To speed up the entire multi-thresholding algorithm and reduce memory requirements, a fractal scanning sub-sampling technique can be used.