Integrated portable shrimp-freshness prediction platform based on ice-templated metal-organic framework colorimetric combinatorics and deep convolutional neural networks

Authors: MA, PEIHUA - UNIVERSITY OF MARYLAND; ZHANG, ZHI - UNIVERSITY OF MARYLAND; XU, WENHAO - UNIVERSITY OF MARYLAND; TENG, ZI - UNIVERSITY OF MARYLAND; Luo, Yaguang - Sunny; GONG, CHENG - UNIVERSITY OF MARYLAND; WANG, QIN - UNIVERSITY OF MARYLAND

Submitted to: ACS Sustainable Chemistry & Engineering
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 11/23/2021
Publication Date: 12/3/2021
Citation: Ma, P., Zhang, Z., Xu, W., Teng, Z., Luo, Y., Gong, C., Wang, Q. 2021. Integrated portable shrimp-freshness prediction platform based on ice-templated metal-organic framework colorimetric combinatorics and deep convolutional neural networks. ACS Sustainable Chemistry & Engineering. 9(50):16926-16936. https://doi.org/10.1021/acssuschemeng.1c04704?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-as.
DOI: https://doi.org/10.1021/acssuschemeng.1c04704?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-as

Interpretive Summary: Real-time and non-destructive monitoring of food quality is critical to reducing food loss and waste. Currently, a significant amount of food products are being discarded at retail stores due to spoilage and/or surpassed the pre-designated expiration dates. To address this issue, we developed a novel sensing platform for predicting food quality by identifying the volatile aromatic compounds released from the products. Via a combination of nanotechnology and artificial intelligence, the sensor monitors signature odor of the food products associated with freshness and spoilage in real-time and non-destructively, providing important information on the extent of product degradation in supply chain. Research findings benefit consumers and the food industry by advancing technologies towards accurate product quality monitoring.

Technical Abstract: Real-time monitoring of the quality and freshness of food products is critical to reducing food loss and waste. Cross-reactive artificial scent screening systems provide a promising solution for freshness monitoring, but their commercialization is hindered by the low detection sensitivity and pattern-recognition inaccuracy. Leveraging cutting-edge artificial intelligence and high-porosity nanomaterial, we developed a low cost and versatile method by incorporating metal-organic frameworks (MOFs) into smart food packaging via a colorimetric combinatorics sensor array. We screened the whole UiO-66 family by density functional theory (DFT) calculations and chose UiO-66-Br to construct sensor arrays on an ice-templated chitosan substrate. The physicochemical properties and morphologies of the fabricated sensor arrays were systematically characterized. The limit of detection (LOD) of 80 mg/L for ammonia was achieved by the prepared composite film. Using shrimp as a food model, deep convolutional neural networks (DCNN) were applied to monitor shrimp freshness by recognizing the scent fingerprint. Four state-of-the-art DCNN models were trained using 31,584 labeled images and 13,537 images for testing. The highest accuracy was achieved (up to 99.94%) by the WISeR50 network. Our newly developed platform is integrated, sensitive, and non-destructive, enabling consumers to conveniently monitor freshness in real-time.