Joint Source-Channel Coding for Wireless Image Transmission: A Deep Compressed-Sensing Based Method

TL;DR

The paper addresses robust wireless image transmission under limited bandwidth by proposing a deep joint source-channel coding framework that embeds block-based compressed sensing (CS) into a CNN-based encoder to produce complex-valued channel inputs. The system trains end-to-end with a non-trainable AWGN channel layer and a CNN-based decoder plus a deep reconstruction network, optimizing a mean-squared error loss under a power constraint $\frac{1}{k} E[z z^*] \\le P$. Key contributions include learning the CS sampling matrix, enforcing the power constraint during normalization $z= \\sqrt{kP} \\,\\frac{\\tilde{z}}{\\sqrt{\\tilde{z}^* \\tilde{z}}}$, and demonstrating substantial PSNR/SSIM gains over DJSCC and ADJSCC on CIFAR-10 and Kodak, with further validation on higher-resolution data via Imagenet-based training. The results indicate that integrating CNN-based CS with DL-based JSCC yields robust, high-fidelity image transmission across diverse SNRs and bandwidths, suggesting practical impact for bandwidth-constrained wireless imaging.

Abstract

Nowadays, the demand for image transmission over wireless networks has surged significantly. To meet the need for swift delivery of high-quality images through time-varying channels with limited bandwidth, the development of efficient transmission strategies and techniques for preserving image quality is of importance. This paper introduces an innovative approach to Joint Source-Channel Coding (JSCC) tailored for wireless image transmission. It capitalizes on the power of Compressed Sensing (CS) to achieve superior compression and resilience to channel noise. In this method, the process begins with the compression of images using a block-based CS technique implemented through a Convolutional Neural Network (CNN) structure. Subsequently, the images are encoded by directly mapping image blocks to complex-valued channel input symbols. Upon reception, the data is decoded to recover the channel-encoded information, effectively removing the noise introduced during transmission. To finalize the process, a novel CNN-based reconstruction network is employed to restore the original image from the channel-decoded data. The performance of the proposed method is assessed using the CIFAR-10 and Kodak datasets. The results illustrate a substantial improvement over existing JSCC frameworks when assessed in terms of metrics such as Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity Index (SSIM) across various channel Signal-to-Noise Ratios (SNRs) and channel bandwidth values. These findings underscore the potential of harnessing CNN-based CS for the development of deep JSCC algorithms tailored for wireless image transmission.

Figures (6)

• Figure 1: Components of the system model R7
• Figure 2: Architecture of the proposed model
• Figure 3: CIFAR-10 dataset: performance of the proposed method versus varying compression ratios over an AWGN channel (PM=Proposed Method)
• Figure 4: CIFAR-10 dataset: performance of different methods with compression ratio $1/6$, versus varying channel SNRs over an AWGN channel (PM=Proposed Method)
• Figure 5: Kodak dataset: performance of different methods with compression ratio $1/6$, versus varying channel SNRs over an AWGN channel (PM=Proposed Method)