Research Article
Image Security Acquisition and Efficient Transmission Algorithm Based on Deep Learning and Neural Network
@ARTICLE{10.4108/eetsis.8413, author={Jianwei Ma and Jing Luo and Zhongqiang Zhou and Yusong Huang and Ling Liang and Chan Wang and Zhencheng Li}, title={Image Security Acquisition and Efficient Transmission Algorithm Based on Deep Learning and Neural Network}, journal={EAI Endorsed Transactions on Scalable Information Systems}, volume={12}, number={6}, publisher={EAI}, journal_a={SIS}, year={2026}, month={1}, keywords={Image Encryption, Generative Adversarial Network, Convolutional Neural Network, Image Compression, Image Restoration}, doi={10.4108/eetsis.8413} }- Jianwei Ma
Jing Luo
Zhongqiang Zhou
Yusong Huang
Ling Liang
Chan Wang
Zhencheng Li
Year: 2026
Image Security Acquisition and Efficient Transmission Algorithm Based on Deep Learning and Neural Network
SIS
EAI
DOI: 10.4108/eetsis.8413
Abstract
INTRODUCTION: Image encryption algorithms of a traditional nature exhibit high computational complexity which in turn creates bottlenecks in performance due to encrypted image operations in real-time image acquisition systems, adversely impacting real-time performance as well as processing efficiency. OBJECTIVES: To this end, this paper applies an image security acquisition and efficient transmission algorithm based on GAN (Generative Adversarial Network) and CNN (Convolutional Neural Network). METHODS: First, a GAN is used for image encryption. By training the generator and discriminator, the generator encrypts the image into an invisible form, and the discriminator ensures that the encrypted image is significantly different from the original image, thereby enhancing the image security. Secondly, CNN is used for image compression. By designing an autoencoder structure, CNN extracts high-level features of the image and compresses it, which reduces bandwidth requirements while ensuring image quality. RESULT: For packet loss or noise pollution that may occur during transmission, the CNN-based image restoration network effectively repairs the missing image part, and the restoration process improves the image restoration quality through multi-level feature extraction and reconstruction technology. CONCLUSION: Experiments show that the model has good real-time performance for large-size images; the SSIM (Structural Similarity Index) is higher than 0.9 in packet loss environments; the transmission delay is less than 0.5 seconds under different compression ratios.
Copyright © 2026 Jianwei Ma et al., licensed to EAI. This is an open access article distributed under the terms of the CC BY-NC-SA 4.0, which permits copying, redistributing, remixing, transforming, and building upon the material in any medium so long as the original work is properly cited.
