Designing Interpretable Recurrent Neural Networks for Video Reconstruction Via Deep Unfolding

Onderzoeksoutput: Articlepeer review

10 Citaten (Scopus)
92 Downloads (Pure)

Samenvatting

Deep unfolding methods design deep neural networks as learned variations of optimization algorithms through the unrolling of their iterations. These networks have been shown to achieve faster convergence and higher accuracy than the original optimization methods. In this line of research, this paper presents novel interpretable deep recurrent neural networks (RNNs), designed by the unfolding of iterative algorithms that solve the task of sequential signal reconstruction (in particular, video reconstruction). The proposed networks are designed by accounting that video frames’ patches have a sparse representation and the temporal difference between consecutive representations is also sparse. Specifically, we design an interpretable deep RNN (coined reweighted-RNN) by unrolling the iterations of a proximal method that solves a reweighted version of the ℓ1-ℓ1 minimization problem. Due to the underlying minimization model, our reweighted-RNN has a different thresholding function (alias, different activation function) for each hidden unit in each layer. In this way, it has higher network expressivity than existing deep unfolding RNN models. We also present the derivative ℓ1-ℓ1-RNN model, which is obtained by unfolding a proximal method for the ℓ1-ℓ1 minimization problem. We apply the proposed interpretable RNNs to the task of video frame reconstruction from low-dimensional measurements, that is, sequential video frame reconstruction. The experimental results on various datasets demonstrate that the proposed deep RNNs outperform various RNN models.
Originele taal-2English
Artikelnummer9394770
Pagina's (van-tot)4099 - 4113
Aantal pagina's15
TijdschriftIEEE Transactions on Image Processing
Volume30
DOI's
StatusPublished - 2 apr 2021

Bibliografische nota

Funding Information:
Manuscript received May 1, 2020; revised January 13, 2021 and March 16, 2021; accepted March 17, 2021. Date of publication April 2, 2021; date of current version April 9, 2021. This work was supported in part by the FWO Research Project under Grant G093817N, in part by the Ph.D. Fellowship Strategic Basic Research under Grant 1SB5721N, and in part by the Flemish Government through the “Onderzoeksprogramma Artifi-ciële Intelligentie (AI) Vlaanderen” Programme. This article was presented at the 2019 IEEE International Conference on Image Processing (ICIP) [1]. The associate editor coordinating the review of this manuscript and approving it for publication was Prof. Denis Kouame. (Corresponding author: Nikos Deligiannis.) The authors are with the Department of Electronics and Informatics (ETRO), Vrije Universiteit Brussel, B-1050 Brussels, Belgium, and also with imec, B-3001 Leuven, Belgium (e-mail: [email protected]; bjoukovs@ etrovub.be; [email protected]). Digital Object Identifier 10.1109/TIP.2021.3069296

Publisher Copyright:
© 1992-2012 IEEE.

Copyright:
Copyright 2021 Elsevier B.V., All rights reserved.

Vingerafdruk

Duik in de onderzoeksthema's van 'Designing Interpretable Recurrent Neural Networks for Video Reconstruction Via Deep Unfolding'. Samen vormen ze een unieke vingerafdruk.

Citeer dit