数字图像的质量可以说在人们进行交流通信和视觉感知的过程中起着至关重要的作用。在数字图像的采集、存储、传输和处理过程中，由于拍摄设备及人为操作不当等诸多因素不可避免使得最终获得的图像失真，或称为图像退化，从而影响图像质量。因此，利用退化图像对原始真实图像进行复原的图像重建技术，一直以来都是数字图像处理领域里的研究热点，对其研究具有十分重要的理论和现实意义。为了获得高质量且高效率的图像重建算法，本项目将研究有机地结合目前主流的两大类不同的图像重建算法：基于优化的图像重建算法和基于深度网络的图像重建算法，充分利用两者的优点，突破已有将深度神经网络当作是一个“黑盒子”的瓶颈和限制，自适应学习图像稀疏表示先验模型和正则化参数，设计具有可解释性和结构性的优化算法启发下的通用图像重建深度神经网络框架。本项目可取得理论创新与技术突破，为新一代的图像重建技术的研发起到积极推动作用。

The quality of digital images plays a significant role in the process of visual perception and communication. The degradations of digital images are inevitably caused by many factors during the period of image acquisition, storage, transmission and processing. Therefore, image reconstruction technology, which infers the recovery of the original image from the observed degraded version, has seen a hot and basic topic in the field of image processing. With the aim of developing a fast yet accurate algorithm for image reconstruction, we combine in this paper the merits of two existing categories of methods: the structure insights of traditional optimization-based methods and the performance/speed of recent network-based ones. This project will study on optimization-inspired interpretable deep neural networks for image reconstruction, with the purpose of surmounting the bottlenecks and constraints of the existing deep neural network as a "black box". All the parameters in the proposed deep networks (e.g. nonlinear transforms, shrinkage thresholds, step sizes, etc.) are learned end-to-end, rather than being hand-crafted. This project will make theoretical innovations and technological breakthroughs, and will further greatly promote the development of the new generation of image reconstruction technology.