An Unpackaged Goods Recognizing Method Based on an Improved YOLOv5 Model

The digital economy and the liberation of labor are important directions of modernization. Combining artificial intelligence (AI) with commercial activity, designing and realizing an intelligent merchandise identification system is a pressing need of the times. The system is based on the improved IM-YOLOv5 (yolov5s+sim+Hardswish) deep training model, which improves the recognition rate of the traditional yolov5 model, and the data shows that the IM-YOLOv5 model improves the recognition accuracy by 5.26% compared to the yolov5 model. Users can be informed of the types of goods in real time through the device, realizing the intelligent transformation of visual recognition commercialization, portability and universality.