Run length coding based new approach to automatic image thresholding

Image thresholding (or binarization) is a basic type of image segmentation capability. To represent binary image objects, run length coding (RL) is frequently utilized. The RL-based adaptive approach is used as a binarization method to obtain the satisfied image intensity discontinuity. Having obtaining the adjustable RL, a multilayer network based on fuzzy inference mechanism is utilized to determine the optimal threshold by use of the global feature measures of the original image. The implementation results show that very visually satisfied results are obtained by the authors´ approach. Comparison with other thresholding methods shows that the generalized RL method is able to determine visually satisfied threshold, to reserve the desired detail texture of original image, and is insensitive to various noises in original images