Related Experiment Videos
A dual-module enhanced YOLO model integrating CSP-PTB for non-destructive freshness grading of Pacific white shrimp
Zhenxing Tian1, Yanyan Wu2, Huan Xiang3
1Key Laboratory of Aquatic Product Processing, Ministry of Agriculture and Rural Affairs, National R&D Center for Aquatic Product Processing, South China Sea Fisheries Research Institute, Chinese Academy of Fishery Sciences, Guangzhou, 510300, China; College of Food Science and Engineering, Ocean University of China, Qingdao, 266000, China.
Abstract:
Peeled Litopenaeus vannamei is highly perishable during storage and transportation, yet current freshness assessment relies on subjective sensory evaluation or time-consuming destructive tests. We propose YOLO_FreshShrimp, an improved YOLOv8-based deep learning model for intelligent freshness grading. High-resolution shrimp images were paired with physicochemical indicators during 4 °C storage to establish visual-freshness mapping. A dataset with four grades (I, II, III, inedible) was constructed. The model integrates a cross-stage partial partially transformer block (CSP-PTB) feature module and a lightweight shared convolutional separate BN detection head (LSCSBD) to address subtle spoilage features and background interference. Ablation studies verified both modules' individual contributions to performance gains. This work delivers a practical, non-destructive, and rapid solution for automated shrimp freshness grading, offering strong potential for industrial quality control applications.
Related Concept Videos
Marine Microbial Ecology
Methods of Controlling Food Spoilage
Deep Sea Microbial Ecology
Differential Staining Technique
Insensitive Nuclei Enhanced by Polarization Transfer (INEPT)
Pore Size Distribution
Adequate...