Target detection method based on feature enhancement and IoU perception

The invention relates to a target detection method based on feature enhancement and IoU perception, and belongs to the field of computer vision target detection. According to the method, the target classification and positioning precision is improved by utilizing the spatial information of the convolutional feature map in the RoI classification regression network, the background information in the RoI features is inhibited and the semantic information in the RoI features is enhanced by utilizing the attention mechanism, the correlation between the category score and the bounding box confidence is increased by utilizing the IoU re-scoring strategy, and the high-quality bounding box is reserved. According to the method, through the RoI classification regression branch network, the spatial information in the feature map can be effectively utilized, and the classification and positioning capabilities of the target detection model are effectively improved; the features in the RoI classification regression branch n