Adaptive and Cascaded Compressive Sensing
This work addresses the challenge of enhancing sampling efficiency in compressive sensing for applications like imaging, though it appears incremental as it builds on existing adaptive CS approaches.
The paper tackles the problem of designing scene-dependent adaptive compressive sensing without ground truth access by proposing an error clamping method based on RIP conditions to predict reconstruction error and adaptively allocate samples, and a cascaded feature fusion network to utilize information from different sampling stages, achieving improved performance over state-of-the-art methods.
Scene-dependent adaptive compressive sensing (CS) has been a long pursuing goal which has huge potential in significantly improving the performance of CS. However, without accessing to the ground truth image, how to design the scene-dependent adaptive strategy is still an open-problem and the improvement in sampling efficiency is still quite limited. In this paper, a restricted isometry property (RIP) condition based error clamping is proposed, which could directly predict the reconstruction error, i.e. the difference between the currently-stage reconstructed image and the ground truth image, and adaptively allocate samples to different regions at the successive sampling stage. Furthermore, we propose a cascaded feature fusion reconstruction network that could efficiently utilize the information derived from different adaptive sampling stages. The effectiveness of the proposed adaptive and cascaded CS method is demonstrated with extensive quantitative and qualitative results, compared with the state-of-the-art CS algorithms.