A Super-resolution Reconstruction Method for Single-frame Character Images Based on Wavelet Neural Network

Abstract

The current single-frame character image reconstruction methods have some problems, such as the noise can not be removed effectively, the anti-interference performance is poor, and the resolution of reconstructed single frame character image is low. In order to solve these problems, this paper proposes a super-resolution reconstruction method of single frame character image based on wavelet neural network. Firstly, analyze the factors affecting image degradation and construct the image degradation model. Secondly, remove the noise in the image using the wavelet threshold denoising method. Then, the image super-resolution reconstruction is completed by wavelet neural network reflection model to improve the resolution of the image. Finally, test the reconstructed single frame character image using MATLAB software. The test results show that the proposed method can effectively remove the image’s noise, and its anti-interference performance is high and the reconstructed image’s resolution is high.