Analysis of Object Detection Model Based on Global Spatial Attention Mechanism
This paper expounds that object detection is a very important and fundamental field of computer vision,which is a visual technique that detects object types in images or videos and labels them with boxes.However,there are difficulties in detecting small targets,identifying occluded or deformed targets.A novel spatial attention module is proposed for this purpose,and a new object detection network is designed based on it,which greatly improves the detection accuracy of the model.The effectiveness of this method has been validated through numerous experiments on the COCO dataset.
NETL
NSTL
万方数据
