An Improved YOLOv10 Algorithm for Kitchen Hygiene and Safety Monitoring

Food hygiene and safety is a critical issue in global public health, particularly in the operational environment of restaurant kitchens where hygiene and safety standards are crucial. With the rapid development of the food industry, traditional manual supervision methods are struggling to cope with the increasingly complex challenges of food hygiene and safety. To address the issue of food hygiene and safety monitoring in kitchen environments, an improved YOLOv10 model is proposed. By incorporating a Global Attention Mechanism (GAM), the model enhances its ability to capture spatial-channel interactions across the entire feature map, improving accuracy in detecting small objects. Additionally, the CIoU loss function is replaced with Wise-IoU (WIoU), which dynamically adjusts gradient gains to reduce the impact of low-quality samples on model training, further enhancing the accuracy and efficiency of kitchen hygiene and safety monitoring. Experimental results show that the improved model achieves high accuracy in complex backgrounds and diverse scenarios, particularly in identifying risk behaviors such as bare skin, unmasked individuals, and rodent activity, with a 2.2 percentage point increase in accuracy compared to the original YOLOv10.