Tracking the freshness of seafood products can be challenging for both consumers and businesses. However, a new study offers a promising solution using artificial intelligence (AI) and traditional science. The use of AI in the shrimp industry is being employed to count and monitor the growth of shrimp.
Researchers from Guangdong Ocean University investigated how the quality of Pacific white shrimp, Litopenaeus vannamei, changes during storage at refrigerated temperatures (4°C) over one week. They employed YOLOv8 as the foundation for developing the YOLO-Shrimp model for rapid detection of quality and freshness.
The Challenge of Keeping Shrimp Fresh
The high protein and moisture content of shrimp make them susceptible to deterioration. Naturally occurring enzymes in shrimp can lead to quality decline, including melanosis (blackening) and loss of flavor and nutrients.
Currently, assessing shrimp freshness relies on measuring biochemical markers such as TVB-N (total volatile basic nitrogen) and TVC (total viable count). While accurate, these methods require sample destruction, bulky equipment, and are time-consuming. This limits their practicality for rapid checks.
Freshness Matters
The study investigates how shrimp (specifically, Pacific white shrimp, Litopenaeus vannamei) stored at refrigerated temperatures (4°C) for one week undergo changes in quality. They focus on three key markers:
- Total volatile basic nitrogen (TVB-N): This chemical compound increases as shrimp degrade, indicating a decrease in freshness.
- Total viable count (TVC): Measures the total number of bacteria present; higher counts indicate greater deterioration.
- Melanosis: Refers to the blackening of shrimp shells, a visual indicator of decreased quality.
The research team also explores the relationships between these markers, providing a more comprehensive picture of shrimp freshness.
YOLO-Shrimp: A Powerful Tool for Freshness Assessment
The study goes beyond traditional methods by introducing the YOLO-Shrimp model. This innovative approach is based on the YOLOv8 architecture, a popular deep learning model for object detection. However, YOLO-Shrimp takes it a step further:
- EIOU focal loss function: This enhancement improves the model’s ability to focus on difficult-to-classify shrimp, yielding more accurate results.
- C3X Computing Module: This module refines the feature extraction process, allowing the model to better identify subtle changes in shrimp texture, color, and shape.
These enhancements translate into significant performance improvements compared to YOLOv8:
- Precision (increased by 5.07%)
- Recall (increased by 1.58%)
- F1 Score (increased by 3.25%)
- mAP50 (increased by 2.84%)
Fundamentally, YOLO-Shrimp model assessments were validated against established biochemical, microbiological, and physical freshness indicators. This strong correlation highlights the model’s effectiveness in detecting shrimp quality.
The science behind the technology
Machine learning algorithms “learn” from data to identify patterns and make predictions. Deep learning, a powerful branch of machine learning, excels in image recognition through convolutional neural networks (CNN). The YOLO algorithm is an efficient object detection framework used for rapid and accurate assessment of food quality based on visual signals.
Benefits for consumers and the industry
Studies confirm that YOLO-Shrimp assessments align with traditional methods like TVB-N and TVC. This highlights its effectiveness in detecting freshness. Here’s how YOLO-Shrimp can benefit the fishing industry:
- Enhanced food safety: Early detection of spoilage allows timely removal of potentially hazardous shrimp.
- Improved quality control: Consistent and objective quality checks lead to better product selection and classification.
- Waste reduction: Non-destructive testing minimizes product damage and ensures efficient utilization.
Conclusion
“In conclusion, the YOLO-Shrimp model, a refined iteration of the YOLOv8 architecture with an integrated EIOU focal loss function and C3X computing module, has demonstrated significant advancements in rapid and non-destructive freshness detection in Litopenaeus vannamei,” conclude the scientists.
This research paves the way for innovative technologies that can effectively monitor and maintain shrimp quality. By combining traditional scientific methods with advanced deep learning models, we can ensure consumers enjoy the freshest and most delicious shrimp possible.
The study was funded by the Natural Science Foundation of China, the Special Fund for Scientific and Technological Innovation Strategy of Guangdong Province, the Guangxi Key Research and Development Plan Project, the Fund of Guangdong Provincial Key Laboratory of Intelligent Equipment for South China Sea Marine Ranching, and the Scientific Research Start-Up Funds of Guangdong Ocean University.
Contact
Ouyang Zheng
College of Food Science and Technology, Guangdong Ocean University
Zhanjiang, Guangdong Province, 524088, China
Email: zhengouyang07@163.com
Reference (open access)
Hou, Mingxin and Zhong, Xiaowen and Zheng, Ouyang and Sun, Qinxiu and Liu, Shucheng and Liu, Mingxin, Innovations in Seafood Freshness Quality: Non-Destructive Detection of Freshness in Litopenaeus Vannamei Using the Yolo-Shrimp Model. Available at SSRN: https://ssrn.com/abstract=4850523 or http://dx.doi.org/10.2139/ssrn.4850523
