Distributed compressed video sensing

Abstract

This paper proposes a novel framework called Distributed Compressed Video Sensing (DISCOS) -a solution for Distributed Video Coding (DVC) based on the recently emerging Compressed Sensing theory. The DISCOS framework compressively samples each video frame independently at the encoder. However, it recovers video frames jointly at the decoder by exploiting an interframe sparsity model and by performing sparse recovery with side information. In particular, along with global frame-based measurements, the DISCOS encoder also acquires local block-based measurements for block prediction at the decoder. Our interframe sparsity model mimics state-of-the-art video codecs: the sparsest representation of a block is a linear combination of a few temporal neighboring blocks that are in previously reconstructed frames or in nearby key frames. This model enables a block to be optimally predicted from its local measurements by l1-minimization. The DISCOS decoder also employs a sparse recovery with side information to jointly reconstruct a frame from its global measurements and its local block-based prediction. Simulation results show that the proposed framework outperforms the baseline compressed sensing-based scheme of intraframecoding and intraframe-decoding by 8 − 10dB. Finally, unlike conventional DVC schemes, our DISCOS framework can perform most encoding operations in the analog domain with very low-complexity, making it be a promising candidate for real-time, practical applications where the analog to digital conversion is expensive, e.g., in Terahertz imaging.Index Terms-distributed video coding, Wyner-Ziv coding, compressed sensing, compressive sensing, sparse recovery with decoder side information, structurally random matrices.

A scheme for distributed compressed video sensing based on hypothesis set optimization techniques

“…This framework, proposed by (Do et al 2009), applies sparse constrained block prediction algorithm. The algorithm assumes w i,m is sparse and it adds the 1 norm as the penalty of the optimization problem, obtaining the weight vector by solving (3).…”

“…It also obtains the information of local complex structure in natural images, and fully exploits the redundancy between frames by combining the block matching technique commonly used in video. Because of these advantages mentioned above, the application of BCS in DCVS has been researched considerably (Do et al 2009;Prades-Nebot et al 2009;Fowler et al 2012;Liu et al 2013Liu et al , 2014. Among them, the BCS-based multi-hypothesis (MH-BCS) scheme has gained wide spread attention due to its high recovery quality.…”

“…They are 1 regularization-based DISCOS scheme (Do et al 2009), Tikhonov regularization-based MH-BCS with smooth projected Landweber reconstruction (MH-BCS-SPL) scheme (Fowler et al 2012) and elastic net-based MS-wEnet scheme (Chen et al 2013) respectively. All of these schemes adopt the MH prediction technique to exploit the correlation between the video frames and obtain a relatively high quality performance.…”

A dictionary generation scheme for block-based compressed video sensing

“…However, compared with such kind of pre-specified basis, dictionaries based on learning can provide a more sparse representation for specific signal data, and what's more, an efficient way to exploit the temporal redundancy in video. In [6,7], a novel inter-frame sparsity dictionary was presented, in which a block could be sparsely represented by a linear combination of a few temporal neighboring blocks. Furthermore, a dictionary learning method based on the training samples extracted from previous reconstructed neighboring frames together with the side information was proposed in [8].…”

“…Then the main challenge is how to utilize the temporal/spatial redundancy in compressed video sensing. In this effort, a number of works have recently appeared [4][5][6][7][8]. A differencing method was used to represent the video frames [4].…”