Download PDFOpen PDF in browserA Proposed Rapid Image Processing-Based Method for Fish Freshness Determination on Mobile ApplicationEasyChair Preprint 364913 pages•Date: June 21, 2020AbstractThe quality of food has been becoming the great concern not only in Vietnam but also all over the globe. Quality of fish in terms of fish freshness is therefore highly attracted by researched and industry community. This paper proposes a simple but effective vision-based method to determine fish’s freshness. The method focuses on status of the fish’s eye through its image taken by smartphone. This determination approach is based on the cornea becoming opaque over time after death, which generally changes the intensity values of different parts within the eyes. Particularly, after the pre-processing with contrast and sharpening adjusting, and segmentation with the Canny algorithm, difference of gray scale intensity between the sclera and the iris areas of the fish eye's image is calculated. The value of the difference is compared with a selected threshold for assessing fish freshness level over the time period from the most freshness to the most staleness. The databases were built consisting of ten samples of fresh popular fish from the local market with three different types: Crucian carp, Gray mullet and Siganus guttatus. The result of observable decrease in the difference of intensity over time on different types of fishes has been recorded which proves the feasibility of the method. Glary effect on the quality of taken image as well as the freshness result has been also minimized and well discussed. This research ultimately introduces an application for smartphone expecting to help people determining the freshness of fish and recognizing the fish type. Keyphrases: Fish Freshness, feature extraction, fish eye segmentation, image processing, mobile application
|