Deep neural network target detection method based on multi-scale receptive field feature fusion

Abstract

The invention relates to a deep neural network target detection method based on multi-scale receptive field feature fusion, and the method comprises the steps: collecting a training sample image whichcomprises three RGB channels, and attaching object detection box labels and the category label information of each object; converting image data and label data in the collected image data set into aformat required by training the deep convolutional neural network through preprocessing; designing a deep convolutional neural network structure based on multi-scale receptive field feature fusion; designing a deep neural network structure applied to target detection, determining an input layer and an output layer of the network according to the structure of input and output data during design, determining the number of multi-scale receptive field feature fusion modules and the number of convolution layers in the neural network, and determining the number of times of network training loop iteration and the final convergence condition of the network; and according to the structure of the trained target and model, defining a required loss function, and in the training stage, carrying out regression on the category of the target and the offset of the detection frame.

Description

technical field [0001] The invention relates to the technical field of computer image recognition, in particular to an image target detection method using a deep neural network method. Background technique [0002] Object detection is one of the important topics in the field of computer vision computing. With the development of society and the advancement of technology, the technology of object detection is constantly being fully used in various scenarios to achieve various expected goals, such as unmanned driving, safety monitoring, video surveillance and traffic control. For a large amount of image and video data and changing scenes, it is of great significance to efficiently locate and classify various objects of interest and achieve fast and accurate object detection. [0003] In recent years, deep learning, especially convolutional neural network, has made great progress in the field of computer vision and natural language processing, which has aroused the research int...

AI Technical Summary

Problems solved by technology

[0012] In order to solve the above-mentioned problems in the existing target detection technology, especially the problem of insufficient positioning ability for targets of different scales, the present invention proposes a deep neural network target detection method based on multi-scale receptive field feature fusion to achieve higher detection accuracy. Speed ​​and Positioning Accuracy

Images