Adaptive Adjustment Object Detection Algorithm under Multiple Mechanisms Based on GAN

Target tracking is prone to problems such as target loss and identity jump when the occluded attitude changed. In order solve this phenomenon, paper proposes adaptive adjustment object detection algorithm under multiple mechanisms based on GAN. This introduces a gradient penalty mechanism discriminator uses relative structure reconstruct discriminator, so improve discriminatory ability of discriminator. Then, through feedback mechanism, obtained data fed back generator network in time, genetic used speed up positioning key areas image. Experimental results show that compared with other existing algorithms, can effectively locate distinguish different environments. And, still maintains high resolution. When occluded, it avoid phenomenon jump.