Threshold selection in image segmentation using parametric entropy measures

Image processing techniques and tools are increasingly used for a variety of commercial and strategic applications leading to a constant need for advancement in this cutting-edge area. Image segmentation is a technique in which homogeneous pixel elements are identified in a given image which then can be used for full object identification or clarification tasks. In this paper, we have attempted to use entropy-based methods for improving the segmentation process by identifying threshold needed for segmentation. Parametric entropy offers flexibility to probe the pixel landscape in a better manner and is found to have advantages.