We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
A SUPER-RESOLUTION RECONSTRUCTION METHOD FOR OF SINGLE-FRAME CHARACTER IMAGES BASED ON WAVELET NEURAL NETWORK.
- Authors
Zhubing Hu; Jian Liu
- Abstract
The image super-resolution reconstruction technology means that a low-resolution image acquired by existing hardware devices or network conditions is reconstructed and converted into a high-resolution image using software processing. The current single-frame character image super-resolution reconstruction methods have problems of poor anti-interference performance and low resolution of reconstructed images. This paper proposes a single-frame character image super-resolution reconstruction method based on wavelet neural network. The current method extracts textures of different directions and image smoothness measures against texture characteristics of crisscrossed and diagonal directions of character images. The super-resolution reconstruction of a single frame image is achieved in the framework of maximum posterior probability. This method has low anti-interference performance, low resolution of reconstructed images. It can be seen that the image resolution obtained by the single-frame character image super-resolution reconstruction method based on wavelet neural network is higher than that based on non-negative neighborhood embedding and non-local regularization and that based on SVR.
- Subjects
IMAGE reconstruction algorithms; IMAGE processing software; IMAGE reconstruction; HIGH resolution imaging
- Publication
Academic Journal of Manufacturing Engineering, 2020, Vol 18, Issue 3, p5
- ISSN
1583-7904
- Publication type
Article