Wireless Image Transmission Using Deep Source Channel Coding With Attention Modules

Recent research on joint source channel coding (JSCC) for wireless communications has achieved great success owing to the employment of deep learning (DL). However, the existing work on DL based JSCC usually trains the designed network in a specific signal-to-noise ratio (SNR) and applies the network for the scenario with the target SNR. A number of networks are required to cover the scenario with a broad range of SNRs, which is computational inefficiency (in the training stage) and requires large storage. These shortages hinder the use of DL based JSCC for real wireless scenarios. We propose a novel method called Attention DL based JSCC (ADJSCC) that can deal with different SNRs with a single neural network. This design is inspired by the resource assignment strategy in traditional JSCC, which dynamically adjusts the compression ratio in source coding and the channel coding rates according to the SNR. As a resource allocation scheme, Attention Mechanism allocates computing resources to more critical tasks, which naturally fits for the resource assignment strategy. Instead of applying the resource allocation strategy in traditional JSCC, the ADJSCC uses the channel-wise soft attention to scale features according to the information of SNR. We compare the ADJSCC method with the state-of-the-art DL based JSCC method through extensive experiments to demonstrate its adaptability, robustness and versatility. Compared with the existing methods, the proposed method takes less storage and is more robust in the burst channel.