Performance metrics comparison of various image segmentation methods
Ligin Alphonsa, R. Resmi · 2016
Segmentation divides an image into different regions with each pixels having similar characteristics. Each of the segmented regions should strongly relate to objects in the image or features of interest for segmentation to be meaningful in image processing. Many methods for image segmentation depends on partitioning of an image into several similar regions. The superpixels generation method have gained substantial interest in the last few years. But all the methods does not meet the post processing time requirement. In this paper the algorithm complexity, boundary recall value, computation time, control on regularity and the aspect of post processing are taken as the performance metrics. Various methods for image segmentation are considered and their different performance metrics are compared in this paper.