Improved BEGAN Algorithm Based on Edge Image Processing

Image edge refers to the set of pixels in the image that have strong contrast with the neighboring graylevel. It is the basic feature of an image and often carriesthe most useful information about the image, which provides a valuable and important basis for people to describe or identify the target and interpret the image. Image enhancement technology is to highlight some features in an image and filters out some unnecessary features according to specific needs. Based on the improved BEGAN data enhancement method, this article proposes a new edge blur background conversion algorithm by optimizing the flow chart of the gray image recognition system, which improves the accuracy of image recognition, and analyzes the advantages and disadvantages of the algorithm and the feasibility of practical use. In order to verify the effectiveness of the improved BEGAN data enhancement method in this article, the algorithm simulation experiment is carried out. The test results show that the accuracy of the improved BEGAN data enhancement method is as high as 95.86%, which is about 15% higher than that of the traditional method. In addition, the MSE and RMSE of the improved algorithm can be stabilized at around 4.11 and 0.518, both of which are at a low level. The test results verify the superiority and reliability of the proposed algorithm.