We propose a novel deep neural network, coined DeepFPC-L2, for solving the 1-bit compressed sensing problem. The network is designed by unfolding the iterations of the fixed-point continuation (FPC) algorithm with one-sided L2-norm (FPC-L2). The DeepFPC-L2 method shows higher signal reconstruction accuracy and convergence speed than the traditional FPC-L2 algorithm. Furthermore, we compare its robustness to noise with the previously proposed DeepFPC network—which stemmed from unfolding the FPC-L1 algorithm—for different signal to noise ratio (SNR) and sign-flipped ratio (flip ratio) scenarios. We show that the proposed network has better noise immunity than the previous DeepFPC method. This result indicates that the robustness of a deep-unfolded neural network is related with that of the algorithm it stems from.
Yang, Y, Xiao, P, Liao, B & Deligiannis, N 2020, A Robust Deep Unfolded Network for Sparse Signal Recovery from Noisy Binary Measurements. in 2020 28th European Signal Processing Conference (EUSIPCO)., 9287582, European Signal Processing Conference, IEEE, pp. 56-60, European Signal Processing Conference, 18/01/21. https://doi.org/10.23919/Eusipco47968.2020.9287582
Yang, Y., Xiao, P., Liao, B., & Deligiannis, N. (2020). A Robust Deep Unfolded Network for Sparse Signal Recovery from Noisy Binary Measurements. In 2020 28th European Signal Processing Conference (EUSIPCO) (pp. 56-60). Article 9287582 (European Signal Processing Conference). IEEE. https://doi.org/10.23919/Eusipco47968.2020.9287582
@inproceedings{bbf0eeadda1d44d897aa7e446d5895be,
title = "A Robust Deep Unfolded Network for Sparse Signal Recovery from Noisy Binary Measurements",
abstract = "We propose a novel deep neural network, coined DeepFPC-L2, for solving the 1-bit compressed sensing problem. The network is designed by unfolding the iterations of the fixed-point continuation (FPC) algorithm with one-sided L2-norm (FPC-L2). The DeepFPC-L2 method shows higher signal reconstruction accuracy and convergence speed than the traditional FPC-L2 algorithm. Furthermore, we compare its robustness to noise with the previously proposed DeepFPC network—which stemmed from unfolding the FPC-L1 algorithm—for different signal to noise ratio (SNR) and sign-flipped ratio (flip ratio) scenarios. We show that the proposed network has better noise immunity than the previous DeepFPC method. This result indicates that the robustness of a deep-unfolded neural network is related with that of the algorithm it stems from.",
author = "Yuqing Yang and Peng Xiao and Bin Liao and Nikolaos Deligiannis",
note = "Publisher Copyright: {\textcopyright} 2021 European Signal Processing Conference, EUSIPCO. All rights reserved.; European Signal Processing Conference, EUSIPCO 2020 ; Conference date: 18-01-2021",
year = "2020",
doi = "10.23919/Eusipco47968.2020.9287582",
language = "English",
isbn = "9781728150017",
series = "European Signal Processing Conference",
publisher = "IEEE",
pages = "56--60",
booktitle = "2020 28th European Signal Processing Conference (EUSIPCO)",
url = "https://eusipco2020.org",
}