Abstract
Compressive sensing (CS) is a technique that enables the recovery of sparse signals using fewer measurements than traditional sampling methods. To address the computational challenges of CS reconstruction, our objective is to develop an interpretable and concise neural network model for reconstructing natural images using CS. We achieve this by mapping one step of the iterative shrinkage thresholding algorithm (ISTA) to a deep network block, representing one iteration of ISTA. To enhance learning ability and incorporate structural diversity, we integrate aggregated residual transformations (ResNeXt) and squeeze-and-excitation (SE) mechanisms into the ISTA block. This block serves as a deep equilibrium layer, connected to a semi-tensor product network (STP-Net) for convenient sampling and providing an initial reconstruction. The resulting model, called MsDC-DEQ-Net, exhibits competitive performance compared to state-of-the-art network-based methods. It significantly reduces storage requirements compared to deep unrolling methods, using only one iteration block instead of multiple iterations. Unlike deep unrolling models, MsDC-DEQ-Net can be iteratively used, gradually improving reconstruction accuracy while considering computation trade-offs. Additionally, the model benefits from multi-scale dilated convolutions, further enhancing performance.
Abstract (translated)
压缩感知(CS)是一种使用比传统采样方法更少的测量的技术,以恢复稀疏信号。为了应对CS重建的计算挑战,我们的目标是开发一个可解释且紧凑的神经网络模型,用于使用CS对自然图像进行重建。我们通过将迭代收缩阈值算法(ISTA)的迭代一步映射到深度网络块来实现这一目标,表示ISTA的迭代。为了提高学习能力和包含结构多样性,我们将聚合残差变换(ResNeXt)和压缩和激发(SE)机制集成到ISTA模块中。这个模块充当一个深度平衡层,连接到半张量产品网络(STP-Net),用于方便的采样和提供初始重建。所得到的模型,称为MsDC-DEQ-Net,与基于网络的先进方法相比具有竞争性能。它显著减少了存储需求,只需使用一个迭代块而不是多个迭代块。与深展开模型不同,MsDC-DEQ-Net可以迭代使用,在考虑计算开销的同时,逐渐提高重建准确性。此外,模型还利用多尺度扩散卷积,进一步提高性能。
URL
https://arxiv.org/abs/2401.02884