One-stage target detection method, system and device based on generative adversarial network

Abstract

The invention belongs to the field of artificial intelligence computer vision, particularly relates to a one-stage target detection method, system and device based on a generative adversarial network,which aim to solve the problem that a one-stage target detector high in speed and high in real-time performance is low in recognition precision of small objects, distorted objects and shielded objects. The method comprises the following steps of based on an acquired input image, acquiring a target image corresponding to each target in the input image through a trained target detection network, constructing a target detection network by combining a generative adversarial network based on a Darknet-53 network framework, constructing a loss function based on the Wasserstein distance function, inthe training process, increasing the number of samples through the distortion feature network, the shielding feature network and the super-resolution feature network. On the premise of ensuring the detection efficiency, the object identification precision of distorted objects, objects under different shielding degrees and small objects is greatly improved.

Description

technical field [0001] The invention belongs to the field of artificial intelligence computer vision, and in particular relates to a one-stage target detection method, system and device based on a generative confrontational network. Background technique [0002] With the continuous improvement of hardware computing power, computer vision has developed rapidly. Computer vision and computer intelligent processing have become an important research field. As an important direction of computer vision, object detection has been developed rapidly. There are more and more applications of target detection algorithms in life scenes, and the fields with a wide range of uses include: unmanned driving, security, logistics sorting, video analysis and other fields. At present, deep learning methods in the field of target detection are mainly divided into two categories: two-stage target detection algorithms and one-stage target detection algorithms. Among them, the two-stage object dete...

AI Technical Summary

Problems solved by technology

[0005] In order to solve the above-mentioned problems in the prior art, that is, the problem of low recognition accuracy of small objects, distorted objects, and occluded objects by a one-stage target detector with high speed and strong real-time performance, the present invention provides a method based on a generative adversarial network A one-stage target detection method, the target detection method includes:

Images