Part pose recognition algorithm based on lightweight CNN
Aiming at problems of current recognition and detection algorithms based on convolutional neural network(CNN),such as a large number of parameters,a huge amount of calculation,a large amount of memory and excessive resource consumption,a lightweight recognition and detection network based on the advantages of the YOLOv3,network structure,EfficientNet-B0-YOLOv3 is proposed.This network can not only realize pose recognition of part,but also can recognize each face of the part.It has high recognition and detection precision,while reducing the amount of parameters and computation of the network,and the size of the trained network model is only 41.10 MB,which can reduce resource consumption.In industrial applications,it reduces memory usage and makes it easier to be embedded devices for use.
万方数据
维普
