D$^3$U-Net: Dual-Domain Collaborative Optimization Deep Unfolding Network for Image Compressive Sensing

Kai Han, Jin Wang, Yunhui Shi, Nam Ling, Baocai Yin

Abstract: Deep unfolding network (DUN) is a powerful technique for image compressive sensing that bridges the gap between optimization methods and deep networks. However, DUNs usually rely heavily on single-domain information, overlooking the inter-domain dependencies. Therefore, such DUNs often face the following challenges: 1) information loss due to the inefficient representation within a single domain, and 2) limited robustness due to the absence of inter-domain dependencies. To overcome these challenges, we propose a deep unfolding framework D$^3$U-Net that establishes a dual-domain collaborative optimization scheme. This framework introduces both visual representations from the image domain and multi-resolution analysis provided by the wavelet domain. Such dual-domain representations constrain the feasible region within the solution space more accurately. Specifically, we design a consistency-difference collaborative mechanism to capture inter-domain dependencies effectively. This mechanism not only enhances the fidelity of reconstruction but also enriches the depth and breadth of extracted features, improving the overall robustness and reconstruction quality. Moreover, we develop an inter-stage transmission pathway to minimize the information loss during transmission while broadcasting multi-scale features in a frequency-adaptive manner. Extensive experimental results on various benchmark datasets show the superior performance of our method.

Primary Subject Area: [Experience] Multimedia Applications

Relevance To Conference: As image is one of the most important forms of multimedia, image reconstruction is the core topic in multimedia applications. Compressed sensing technology introduces a novel approach by reducing the number of samples required during the acquisition process while maintaining the quality and reconstructability of the signal. In this work, we propose a novel framework that establishes a dual-domain collaborative optimization scheme to significantly enhance the quality of reconstructed images. In this framework, both visual representation from the image domain and multi-resolution analysis provided by the wavelet domain are introduced to highly improve the visual experience for users. A consistency-difference collaborative mechanism is designed to effectively capture inter-domain dependencies. Through experiments, we have demonstrated the effectiveness of this method in improving image quality, reducing reconstruction errors, and enhancing the visual experience.

Submission Number: 4521