Computer vision based method for quality and freshness check for fish from segmented gills

The quality and freshness of a fish sample is principally hampered in the post-harvested phase due to storage, handling and processing. The quality of the sample may degrade as the days pass till it finally reaches the consumers. The quality of a post harvested fish is determined mainly by two important factors namely climatic conditions and holding time. This paper presents a completely automated computer vision based segmentation of fish gills from digital images of fish samples. Post segmentation, a statistical relationship of the segmented gill region is established to design an assessment model for fish freshness identification. The fish gills are segmented using various strategic image processing techniques like contrast enhancement, adaptive intensity threshold and active contour based methods. The model for fish freshness testing is based on the image statistical features which are derived from the gills region of the saturation channel. The variation of the statistical distribution is observed to be decreasing monotonic which is basis for design of the framework for fish quality and freshness identification. This process being non-destructive provides an efficient fish quality assessment scheme in real time. (C) 2017 Elsevier B.V. All rights reserved.