A novel pilot study on imaging‐based identification of fish exposed to heavy metal (Hg) contamination

Abstract

Heavy metal pollution in water negatively affects the health status of fishes and makes it toxic to human health. Conventional chemical-based methods for fish quality assessment are destructive, highly time-consuming, and require expensive machines and expert manpower. In this novel pilot study, the image of the gill tissue of fish is considered as the main region of interest to analyze the discriminatory behavior between normal and heavy-metal exposed fish to develop an automatic and nondestructive approach using computer vision and machine learning algorithms. The discriminatory behavior of extracted features was analyzed with the help of p-value criteria, and the proposed framework is found efficient to identify heavy metal exposed fish where the value of the area is in the range of 82% to 92% under the curve of the receiver operating characteristic. This paper establishes the scientific proof of concept of using gill tissue of the fish for the identification of heavy metal exposed fish. Novelty impact statement An automatic and non-destructive method is proposed for the identification of heavy metal exposed fish using computer vision and machine learning algorithms. The region of interest is automatically segmented from an image using thresholding and morphological operations. This paper establishes the scientific proof of concept of using gill tissue of the fish for the identification of heavy metal exposed fish.