Research on the Restoration of Sparse Image based on Deep Learning

DOI: https://doi.org/10.62517/jbdc.202301201

Author(s)

Luqi He*, Jiaming Gong, Ning Wang, Yunxin Kuang, Canyao Zhao, Yimei Peng

Affiliation(s)

College of Data Science, Guangzhou Huashang College, Guangzhou, Guangdong, 511300, China *Corresponding Author.

Abstract

This article mainly explores the sparse image information restoration technology based on deep learning. Through in-depth analysis of image damage and noise characteristics, this paper proposes a novel method aimed at restoring sparse information from images with damage and noise. This paper introduces the characteristics of image damage and noise, and proposes a deep learning based solution to address these issues. By using deep neural networks, the model proposed in this article can effectively learn sparse information in images and recover it from damage and noise. This paper elaborates on the implementation process of the proposed model, including core elements such as the design of loss functions and the selection of optimization algorithms. Specifically, the loss function proposed in this article can effectively measure the accuracy of model predictions and help the model better learn sparse information in images. In addition, this paper provides a comprehensive review and outlook on relevant work, summarizes the shortcomings of current research work, and proposes future research directions. In summary, the sparse image information restoration technology based on deep learning proposed in this article can effectively improve image quality and provide new ideas and directions for research in related fields.

Keywords

Deep Learning; Sparse Information; Loss Function; Optimization Algorithm

References

[1] Li Junshan, Yang Yawei, Zhang Jiao, et al. 2018, Research progress on degraded image restoration methods. Liquid Crystal and Display, 33 (08): 676-689 [2] Li Junfei, Xu Liming, Wang Yang, et al. 2023, A review of research on citation classification of scientific and technological literature based on deep learning technology. Frontiers of Data and Computing Development, 5 (04): 86-100 [3] Zhang Yanhong, Qin Fengqing, Jiang Li, et al. 2022, Research on Gaussian Blurred Image Restoration with Constrained Least Squares Filter. Volkswagen Technology, 24 (09): 10-13 [4] Shi Ying, Qi Hui. 2019, An Image Restoration Method Based on LMBP Neural Network. Journal of Taiyuan Normal University (Natural Science Edition), 18 (02): 64-67 [5] Mo Siyu, Zhou Qing. 2018, Image Restoration Method Based on MEA-BP Neural Network. Electronic Design Engineering, 26 (21): 185-189 [6] Li Qingquan, Wang Huan. 2019, Overview of Optimization Algorithms Based on Sparse Representation Theory. Surveying and Mapping Geographic Information, 44 (04): 1-9. [7] Shan G. 2020, The study of signal reconstruction: an investigation independent of Shannon-Nyquist theorems. Science China (Information Sciences), 63(10):279-281. [8] Tao Wenbin, Qian Yurong, Zhang Yiyang, et al. 2022, A review of deep clustering algorithms based on autoencoders. Computer Engineering and Applications, 58 (18): 16-25 [9] Yue X, Junyu W, Lei Y. 2023, Orthogonal matching pursuit algorithm based on weighted cosine similarity for sound source localization. Physica Scripta, 98(9). [10] Zhao Xiaohong, Heng Chunni, Du Shujuan. 2020, Medical image convolutional super-resolution reconstruction for residual networks and Adam optimization algorithm. Adhesive, 44 (12): 78-85