Affiliation:
1. Yangzhou University
2. Institut Supérieur d'Électronique de Paris
3. Polytechnique Montreal
4. Sorbonne University
Abstract
Abstract
Closed-loop architecture is widely utilized in automatic control systems and attains distinguished dynamic and static performance. However, classical compressive sensing systems employ an open-loop architecture with separated sampling and reconstruction units. Therefore, a method of iterative compensation recovery for image compressive sensing (ICRICS) is proposed by introducing a closed-loop framework into traditional compressive sensing systems. The proposed method depends on any existing approaches and upgrades their reconstruction performance by adding a negative feedback structure. Theoretical analysis of the negative feedback of compressive sensing systems is performed. An approximate mathematical proof of the effectiveness of the proposed method is also provided. Simulation experiments on more than 3 image datasets show that the proposed method is superior to 10 competing approaches in reconstruction performance. The maximum increment of the average peak signal-to-noise ratio is 4.36 dB, and the maximum increment of the average structural similarity is 0.034 based on one dataset. The proposed method based on a negative feedback mechanism can efficiently correct the recovery error in the existing image compressive sensing systems.
Publisher
Research Square Platform LLC
Reference64 articles.
1. Y. Yang, J. Sun, H.B. Li, and Z.B. Xu, “DMM-CSNet: A deep learning approach for image compressive sensing,” IEEE Trans PAMI, vol. 42, no. 3, pp. 521–538, Mar. 2020, doi: 10.1109/TPAMI.2018.2883941.
2. W.Z. Shi, F. Jiang, S.H. Liu, A. Teramoto, and D.B. Zhao, “Image compressed sensing using convolutional neural network,” IEEE Trans IP, vol. 29, pp. 375–388, Jan. 2020, doi: 10.1109/TIP.2019.2928136.
3. C.A. Tavares, T.M.R. Santos, N.H.T. Lemes, J.P.C. dos Santos, J.C. Ferreira, and J.P. Braga, “Solving ill-posed problems faster using fractional-order Hopfield neural network,” Journal CAM, vol. 381, pp. 1–14, Jan. 2021, doi: 10.1016/j.cam.2020.112984.
4. On the second-order asymptotical regularization of linear ill-posed inverse problems;Zhang Y;Applicable Analysis,2020
5. J. Adler and O, Oktem, “Solving ill-posed inverse problems using iterative deep neural networks,” Inverse Problems, vol. 33, no. 12, pp. 1–10, Dec. 2017, doi: 10.1088/1361-6420/aa9581.