Search

Seaweed has attracted great attention as a healthy and nutritious food. Traditional seaweed processing plants mainly rely on manual visual inspection to identify and remove inferior seaweed. Accurate and rapid harvest period classification and impurity detection are key to improving productivity and processing speed in seaweed processing plants. Although many studies on seaweed have been conducted in laboratory environments, currently, the plants lack effective tools to obtain real-time and reliable information on seaweed quality. To address this challenge, the deep transfer learning-based computer vision was applied to identify inferior seaweeds, including those from the third harvest, fourth harvest, and impure seaweeds in this work. Specifically, YOLOv8 and YOLOv5 were utilized as base transfer learning models. By loading various pre-trained weight files, this study was able to automatically classify Porphyra haitnensis into four categories based on the harvest period and simultaneously detect four types of common impurities in it. Among the tested models, YOLOv8n-cls achieved the best trade-off in classifying the harvest period, with a Top-1 accuracy of 93.5%. This represented a significant improvement of 16% compared to the performance without transfer learning. The detection speed for a single image was 8.2 ms, and the model size was only 2.82 Mb. On the other hand, YOLOv8n exhibited outstanding performance in impurity detection, with a mean average precision of 99.14%, a single image detection speed of 4.3 ms, and a model size of 5.95 Mb. The results demonstrated the potential of YOLOv8 with transfer learning to objectively assist or even replace decision-making by assembly line workers. This study will not only enhance the quality control, production efficiency, and economic benefits of the seaweed processing industry but also drive the automation equipment and systems of seaweed-related enterprises towards greater intelligence and efficiency. [ABSTRACT FROM AUTHOR]

Hydrocolloids are widely used in meat products as common food additives. However, research has indicated that excessive consumption of these hydrocolloids may have potential health implications. Currently, consumers mainly rely on sensory evaluation to identify hydrocolloid adulteration in meat products. Although many studies on quantitative detection of hydrocolloids have been conducted by biochemical methods in laboratory environments, there is currently a lack of effective tools for consumers and regulators to obtain real‐time and reliable information on hydrocolloid adulteration. To address this challenge, a smartphone‐based computer vision method was developed to quantitatively detect carrageenan adulteration in beef in this work. Specifically, Swin Transformer models, along with pre‐training and fine‐tuning techniques, were used to successfully automate the classification of beef into nine different levels of carrageenan adulteration, ranging from 0% to 20%. Among the tested models, Swin‐Tiny (Swin‐T) achieved the highest trade‐off performance, with a Top‐1 accuracy of 0.997, a detection speed of 3.2 ms, and a model size of 103.45 Mb. Compared to computer vision, the electrochemical impedance spectroscopy achieved a lower accuracy of 0.792 and required a constant temperature environment and a waiting time of around 30 min for data stabilization. In addition, Swin‐T model was also capable of distinguishing between different types of hydrocolloids with a Top‐1 accuracy of 0.975. This study provides consumers and regulators with a valuable tool to obtain real‐time quantitative information about meat adulteration anytime, anywhere. Practical Application: This research provides a practical solution for regulators and consumers to non‐destructively and quantitatively detect the content and type of hydrocolloids in beef in real‐time using smartphones. This innovation has the potential to significantly reduce the costs associated with meat quality testing, such as the use of chemical reagents and expensive instruments. [ABSTRACT FROM AUTHOR]