A low-cost imaging framework for freshness evaluation from multifocal fish tissues. (February 2022)

• Record Type: Journal Article
• Authors: Arora, Monika; Mangipudi, Parthasarathi; Dutta, Malay Kishore
• Is Part Of: Journal of food engineering. Volume 314(2022)
• Publication Date: 2022-02
• Subjects: Computer vision -- Feature extraction -- Freshness coefficient -- Multifocal fish tissues -- Weighted features -- Q-score; Food industry and trade -- Periodicals; Food -- Analysis -- Periodicals; Aliments -- Industrie et commerce -- Périodiques; Aliments -- Analyse -- Périodiques; Aliments -- Recherche -- Périodiques; 664.005
• DOI: 10.1016/j.jfoodeng.2021.110777
• Languages: English
• ISSNs: 0260-8774

Abstract

Abstract: Fish is one of the most nutritive food products whose quality gets affected along the food supply chain from harvesting to consumption. Freshness of fish, during that time, gets affected owing to the chemical decomposition of focal tissues like gills, eyes and skin. A novel mathematical model, is being proposed in this article, for the computation of a distinct freshness coefficient, Q-score, obtained by fusing the information from relevant focal tissues like gills, eyes, and skin of the fish. The mathematical model was designed by employing a meta-heuristic approach, where computationally optimized weights have been assigned to the features derived from the focal tissues. These multifocal tissues have been thoroughly investigated and assertive efforts have been made to normalize and integrate the extracted features into a single score. Implementation of the framework, provides an accuracy of 98.07%, indicating the efficacy of the proposed work. The proposed method is automatic and rapid for food quality evaluation. Imaging-based framework for the identification of fish freshness makes the proposed approach non-destructive which upholds its capability to be used in real-time situations. Highlights: Identification of significant focal tissues during the chemical decomposition of fish. Utilization of multiple and prominent focal tissues such as gills, eyes and skin. Estimating weights on tissues and extracted features based upon its discrimination. Computing freshness coefficient (Q-score), obtained from integrating feature weights. Building a model which is efficient and robust to illumination and imaging artifacts.