SuperTA-Net: A Supervised Transmission-Augmented Network for Image Compressive Sensing Reconstruction

Compressive sensing (CS) has seen extensive use in signal processing, particularly in tasks related to image reconstruction. CS simplifies the sampling and compression procedures, but leaves the difficulty to the nolinear reconstruction. Traditional CS reconstruction algorithms are usually iterative, having a complete theoretical foundation. However, these iterative algorithms are constrained by significant computational complexity. While modern deep network-based methods can achieve high-precision reconstruction in compressed sensing (CS) with satisfactory speed, they often lack theoretical analysis and interpretability. To leverage the strengths of both types of CS methods, the deep unfolding networks (DUNs) have been developed. In this paper, a novel DUN named supervised transmission-augmented network (SuperTA-Net) is proposed. Based on the framework of our previous work PIPO-Net, the multi-channel transmission strategy is put forward to reduce the influence of critical information loss between modules and improve the reliability of data. Besides, in order to avoid the issues such as high information redundancy and high computational burden when too many channels are set, the attention based supervision scheme is presented to dynamically adjust the weight of each channel and remove the redundant information. Through experiments focused on reconstructing CS images, the proposed neural network architectures are shown to be highly effective.