Research on Image Multi-target Feature Recognition Technology Based on CNN Network
The existing image feature recognition algorithms have problems in the recognition process,such as single fea-ture points,low feature recognition accuracy,and single recognition target features.Especially when recognizing multiple target features in the image simultaneously,there are problems such as target feature misrecognition or even inability to recognize.In response to the above issues,this paper takes the multi-objective feature requirements of image data as the driving force,and combines the image feature recognition process to carry out business process design.It also carries out the construction of an image multi-objective feature recognition system based on CNN network,and synchronously designs functional units such as image filtering function,image target feature localization,image feature enhancement,and CNN neural network.
NETL
NSTL
万方数据
