Patent
Variable step size distributed compressed sensing reconstruction method based on recurrent neural network
Zeng Chunyan,Wu Minghu,Wan Xiangkui,Xiong Wei,Liu Min,Zhao Nan,Zhu Li,Li Lirong,Wang Juan,Rao Zheheng +9 more
- 03 Nov 2017
3
TL;DR: In this article, a variable step size distributed compressed sensing reconstruction method based on a recurrent neural network is proposed, which comprises the following steps: acquiring structural information of a vector to be reconstructed of each channel by using the RNN, obtaining each non-zero conditional probability of the vector in each channel, then estimating an optimal atom of the current iteration, and then determining a value of a nonzero term of a channel by solving a least square problem, and completing the reconstruction of signals.
read more
Abstract: The invention belongs to the field of distributed compressed sensing reconstruction technology, and particularly relates to a variable step size distributed compressed sensing reconstruction method based on a recurrent neural network. The method comprises the following steps: acquiring structural information of a vector to be reconstructed of each channel by using the recurrent neural network, obtaining each non-zero conditional probability of the vector in each channel, then estimating an optimal atom of the current iteration, and then determining a value of a non-zero term of each channel by solving a least square problem, and completing the reconstruction of signals. According to the method, non-combined sparse multi-channel signals can be reconstructed, and meanwhile, the computational complexity of the coding end cannot be increased.
read more
Chat with Paper
AI Agents for this Paper
Find similar papers on Google Scholar, PubMed and Arxiv
Write a critical review of this paper
Analyze citations of this paper to find unaddressed research gaps
Citations
Patent
A neural network pruning method based on probability
Wang Huan,Hu Haoji,Wang Yuehai +2 more
- 18 Dec 2018
TL;DR: In this paper, a neural network pruning method based on probability is proposed to solve the problems of large storage amount and large calculation amount of the depth learning model represented by the convolution neural network.
2
Patent
Voice transmission method, voice transmission system and terminal
Liu Xiaodong,Meng Fanjing,Li Mingjing +2 more
- 14 Dec 2018
TL;DR: In this article, the authors combined compressed sensing and the recurrent neural network RNN technology in the voice transmission process, and the original voice signal is preprocessed into the time sequence signal S at the transmitting end, the main feature signal is extracted by the compressed sensing technology and then the extracted main signal is coded by the RNN to form the voice coding sequence H.
Patent
Video compressed sensing reconstruction method based on LSTM network and image group quality blind evaluation
Liu Hao,Wei Dong,Zhou Jian,Tian Wei,Chen Genlong,Huang Rong,Sun Shaoyuan,Demin Li,Zhou Wuneng,Wei Guolin,Liao Rongsheng,Huang Zhen +11 more
- 25 Oct 2019
TL;DR: In this paper, a video compressed sensing reconstruction method based on an LSTM network and image group quality blind evaluation is proposed, which comprises: a reconstruction end receiving a frame observation vector code stream; and combining to form continuous image group observation vectors, executing multi-frame joint iterative reconstruction based on the LSTMs on each image group to obtain a corresponding reconstructed image group, outputting final reconstructed frames one by one, and determining whether to update a parameter set of the lstM network according to a continuous condition that the number of iterations reaches a maximum value.
References
Distributed Compressive Sensing: A Deep Learning Approach
TL;DR: The long short-term memory (LSTM) is proposed, a data-driven model for sequence modeling that is deep in time that significantly outperforms the general MMV solver (the Simultaneous Orthogonal Matching Pursuit) and a number of the model-based Bayesian methods.
Patent
Variable step size regularized adaptive compressed sampling matching pursuit method
Yong Liao,Zhou Xin,Li Yufeng,Chen Min An,Chen Ling,Zhang Shumin +5 more
- 27 Jan 2016
TL;DR: In this paper, a variable step size regularized adaptive compressed sampling matching pursuit (SAMP) method is proposed to solve the problem of sparseness K is difficult to obtain in reality, and the adaptive idea of the sparse adaptive matching pursuit is applied in ROMP.
13
Patent
Multiple observed value vector sparsity self-adaptive compressed sampling matching pursuit method
Jia Min,Shi Yao,Gu Xuemai,Guo Qing,Liu Xiaofeng,Wang Xue,Chen Ziyan,Zhu Siyu +7 more
- 27 Jan 2016
TL;DR: In this article, a multiple observed value vector sparsity self-adaptive compressed sampling matching pursuit method is proposed to solve the problem of recovering an original multiband signal from multiple observations with unknown sparsity after continuous-limited module conversion through sampling by a modulated broadband converter under an Xampling framework.
7
Patent
Compressed sensing based wireless sensor image capturing and transmitting system
Hu Chunhai,Liu Hongxiang,Liu Yanling,Liu Bin +3 more
- 23 Mar 2016
TL;DR: In this paper, a compressed sensing based wireless sensor image capturing and transmitting system is proposed, which consists of a terminal node, an aggregation node and a private computer (PC), and the system can form a sensor network through self-organization, so as to automatically capture, compress and send images in a detection region.
5