In-Depth Investigation of Different Segmentation Techniques on Local Fish Images

Abstract

Over the year’s pictures are one of the major information sharing outlets. Nowadays Image segmentation is used to identifying the objects as well as boundaries in the images. Due to the advancement of computer vision technology I try to investigate different image segmentation techniques on local fish images. There are several algorithms proposed for segmenting an image. In this paper five segmentation techniques for local fishes are discussed, they are Otsu Method for Thresholding Algorithm, Histogram based Algorithm, K-means Segmentation Algorithm, Edge Detection Algorithm, Region Growing Algorithm. The results revealed that the Otsu Method for Thresholding Algorithm has achieved a very good result comparing with other techniques. A fish dataset consisting of real-world images was tested. More than 90% accuracy has been achieved, which appears to be good and promising by comparing the performances obtained with the relevant works recently reported.