End-to-End Video Compressive Sensing Using Anderson-Accelerated Unrolled Networks

Abstract

Compressive imaging systems with spatial-temporal encoding can be used to capture and reconstruct fast-moving objects. The imaging quality highly depends on the choice of encoding masks and reconstruction methods. In this paper, we present a new network architecture to jointly design the encoding masks and the reconstruction method for compressive high-frame-rate imaging. Unlike previous works, the proposed method takes full advantage of a denoising prior to provide a promising frame reconstruction. The network is also flexible enough to optimize full-resolution masks and efficient at reconstructing frames. To this end, we develop a new dense network architecture that embeds Anderson acceleration, known from numerical optimization, directly into the neural network architecture. Our experiments show the optimized masks and the dense accelerated network respectively achieve 1.5 dB and 1 dB improvements in PSNR without adding training parameters. The proposed method outperforms other state-of-the-art methods both in simulations and on real hardware. In addition, we set up a coded two-bucket camera for compressive high-frame-rate imaging, which is robust to imaging noise and provides promising results when recovering nearly 1,000 frames per second. Index Terms-high-frame-rate imaging, deep neural network, computational camera ! • This paper is under review for ICCP 2020 and the PAMI special issue on computational photography. Do not distribute. [6]. However, the encoding and decoding components of 35 the imaging system are highly interdependent. Based on 36 this observation, we focus on the joint end-to-end design of 37 encoding masks and reconstruction methods for improving 38 both encoding efficiency and reconstruction accuracy. We 39 put forward a compact end-to-end neural network that can 40 handle the mask optimization for the whole image with 41 fewer training parameters. We also show that this network 49 • We present the first work to jointly design full-reso-50 lution coding masks and reconstruction methods for 51 compressive high-frame-rate imaging using an end-52 to-end network. Our approach outperforms state-of-53 art methods by 2.2dB in PSNR. 54 • We show that the acceleration of the gradient de-55 scent algorithm is equivalent to adding dense skip 56 connections to iterative optimization-unrolling neu-57 ral networks. This speeds up training convergence 58 and helps to design a compact and efficient network 59 architecture. 60 • Experiments on both simulation and real hardware 61 demonstrate the effectiveness of our reconstruction 62 method and the designed masks. The two-bucket 63 design of our camera shows improved noise sup-64 pression and can provide promising results in re-65 constructing video of frame rates up to almost 1,000 66 frames per second. 67 2 RELATED WORK 68 Many approaches have been developed to solve the ill-69 conditioned inverse problem in CS. The existing methods 70 can be divided into model-based optimization methods, 71 deep discriminative learning methods, and unrolled itera-72 t...

Adaptive Deep PnP Algorithm for Video Snapshot Compressive Imaging

“…However, the slow speed of DeSCI precludes its real applications. Even though deep learning methods [6,7,15,20,28,37,38] can achieve the state-of-the-art results within seconds (after training), they lose the robustness of the network for new sensing matrix. By contrast, the previous study [44] has proved that PnP combining deep denoising prior can provide an efficient and flexible method for video SCI reconstruction.…”

“…Among the optimization ones, TwIST [2], Gaussian Mixture Model (GMM) in [40,41] and GAP-TV [16,42] have a high computing speed but could not achieve high quality of images, while DeSCI [17] can obtain high quality images, but the computing process takes a long time. Recently, many deep learning methods have been developed to reconstruct the videos for the SCI system [6,7,15,28,31]. For typical ill-posed problem of SCI, optimization inspired physicsdriven networks (deep unfolding) [20,22,38] have been proposed.…”

Computational Imaging and Artificial Intelligence: The Next Revolution of Mobile Vision

“…With the development of deep learning, more optimization of the speed and logic of the SCI networks are proposed. By embedding Anderson acceleration into the network unit, a deep unrolling algorithm has been developed for SCI reconstruction [181]. Considering the temporal correlation in video frames, the recurrent neural network has also been a tool [182] [183] for video SCI.…”

Two-Stage is Enough: A Concise Deep Unfolding Reconstruction Network for Flexible Video Compressive Sensing

“…• Aiming to extract the temporal correlations among adjacent frames, recurrent neural networks have been used for VCS reconstruction, dubbed BIRNAT [4]. • As a novel approach to combine iterative algorithms and deep learning networks, deep unfolding/unrolling approaches [28,29] have also been used for VCS reconstruction [10,14,25]. Another gain of deep unfolding is the interpretability [35], where the deep networks play the role of denoising in each stage.…”