Adaptive dynamic run-length coding for image segmentation
Sara Arasteh, Chih‐Cheng Hung, Enmin Song · 2006
In this study, an adaptive image segmentation algorithm based on a modified Dynamic Window-based gray-level Run-Length Coding (DW-RLC) is developed. We propose an adaptive method for determining the range which is used in the (DW-RLC) algorithm for image segmentation. First, the modes are estimated and identified from the image histogram automatically. Secondly, the adaptive range (R) is calculated using the statistics estimated from the mode distribution. Then, the actual interval for the segmentation is updated based on the seed pixel value and the adaptive range (R). The proposed adaptive algorithm can be categorized as a region growing method. The proposed method is compared with the mean shift and K-means algorithms. This method improves the DW-RLC algorithm by making it less dependent on the randomly chosen range for segmentation, hence more stable.