Mask-RCNN with Attention Mechanism for Detection and Segmentation

In many existing hospitals and pharmacies, check medicine is done manually which is a time-consuming process. In addition, simple repetitive tasks can lead to eye fatigue and improper drug tests can lead to medical accidents. For the sake of solve this problem, this paper proposes a object detection and semantic segmentation algorithm based on improved Mask R-CNN. We use a new feature extraction network ResNeXt to improve the generalization ability of the feature extraction network. Moreover, the Coordinate Attention mechanism is added to the feature extraction network to increase the attention to location information and improve the performance of the network. Experiments show that the proposed method is superior to the previous methods in object detection and instance segmentation.