Underdetermined blind separation of weak sparse sources via matrix transform layer by layer in the Time-Frequency domain

2015 ◽  
Author(s):  
Xiaohong Ma ◽  
Shuxue Ding ◽  
Jifei Song ◽  
Dongyan Zhu
Keyword(s):  
Frequency Domain ◽  
Layer By Layer ◽  
Blind Separation ◽  
Time Frequency

Blind Source Separation Based on Wavelet and Cross-Wavelet

2011 ◽  
Vol 328-330 ◽  
pp. 2064-2068 ◽  
Author(s):  
Jing Hui Wang ◽  
Yuan Chao Zhao
Keyword(s):  
Frequency Domain ◽  
Blind Source Separation ◽  
Time Scale ◽  
Source Separation ◽  
Blind Separation ◽  
Time Frequency ◽  
Non Stationary Signal ◽  
Cross Wavelet

In this paper, a novel blind separation approach using wavelet and cross-wavelet is presented. This method extends the separate technology from time-frequency domain to time-scale domain. The simulation showed that this method is suitable for dealing with non-stationary signal.

scholarly journals Underdetermined Blind Separation of Nondisjoint Sources in the Time-Frequency Domain

2007 ◽  
Vol 55 (3) ◽  
pp. 897-907 ◽  
Author(s):  
Abdeldjalil Aissa-El-Bey ◽  
Nguyen Linh-Trung ◽  
Karim Abed-Meraim ◽  
Adel Belouchrani ◽  
Yves Grenier
Keyword(s):  
Frequency Domain ◽  
Blind Separation ◽  
Time Frequency

scholarly journals A Statistical Image Feature-Based Deep Belief Network for Fire Detection

Complexity ◽  
2021 ◽  
Vol 2021 ◽  
pp. 1-12
Author(s):  
Dali Sheng ◽  
Jinlian Deng ◽  
Wei Zhang ◽  
Jie Cai ◽  
Weisheng Zhao ◽  
...  
Keyword(s):  
Frequency Domain ◽  
Image Features ◽  
Deep Belief Network ◽  
Image Feature ◽  
Support Vector ◽  
Layer By Layer ◽  
Complex Environment ◽  
Belief Network ◽  
Time Frequency ◽  
Feature Based

Detecting fires is of significance to guarantee the security of buildings and forests. However, it is difficult to fast and accurately detect fire stages in complex environment because of the large variations of the fire features of color, texture, and shapes for flame and smoke images. In this paper, a statistic image feature-based deep belief network (DBN) is proposed for fire detections. Firstly, for each individual image, all the statistic image features extracted from a flame and smoke image in time domain, frequency domain, and time-frequency domain are calculated to construct training and testing samples. Then, the constructed samples are fed into DBN to classify the multiple fire stages in complex environment. DBN can automatically learn fault features layer by layer using restricted Boltzmann machine (RBM). Experiments using the benchmark data of three groups of fire and fire-like images are classified by the present method, and the classification results are also compared with those commonly used support vector machine (SVM) and convolutional deep belief networks (CDBNs) to manifest the superiority of the classification accuracy.

scholarly journals Blind separation of complex-valued satellite-AIS data for marine surveillance: a spatial quadratic time-frequency domain approach

2019 ◽  
Vol 9 (3) ◽  
pp. 1732
Author(s):  
Omar Cherrak ◽  
Hicham Ghennioui ◽  
Nadege Thirion Moreau ◽  
El Hossein Abarkan
Keyword(s):  
Frequency Domain ◽  
Block Diagonalization ◽  
Blind Separation ◽  
Time Frequency ◽  
High Magnitude ◽  
Frequency Domain Methods ◽  
Stationary Signals ◽  
Mixing Matrix ◽  
Frequency Domain Approach ◽  
Complex Valued

<p>In this paper, the problem of the blind separation of complex-valued Satellite-AIS data for marine surveillance is addressed. Due to the specific properties of the sources under consideration: they are cyclo-stationary signals with two close cyclic frequencies, we opt for spatial quadratic time-frequency domain methods. The use of an additional diversity, the time delay, is aimed at making it possible to undo the mixing of signals at the multi-sensor receiver. The suggested method involves three main stages. First, the spatial generalized mean Ambiguity function of the observations across the array is constructed. Second, in the Ambiguity plane, Delay-Doppler regions of high magnitude are determined and Delay-Doppler points of peaky values are selected. Third, the mixing matrix is estimated from these Delay-Doppler regions using our proposed non-unitary joint zero-(block) diagonalization algorithms as to perform separation.</p>

Blind Separation of Non-Stationary Convoluted Mixtures Based on Time-Frequency Analysis

2012 ◽  
Vol 538-541 ◽  
pp. 2571-2575
Author(s):  
Peng Wang ◽  
Ji Hua Cao ◽  
Xiao Chang Ni
Keyword(s):  
Frequency Domain ◽  
Frequency Analysis ◽  
Time Varying ◽  
Blind Separation ◽  
Frequency Method ◽  
Distribution Method ◽  
Time Frequency Analysis ◽  
Joint Diagonalization ◽  
Time Frequency ◽  
Analysis Methods

The signals of convoluted mixtures have a stated of non-stationary identity, and the change of their spectrum with time-varying usually could not be observed from the frequency domain, but they can be observed by the time-frequency method. Therefore, the blind separation of non-stationary convoluted mixtures based on time-frequency analysis is proposed in this paper. For the non-stationary identity, the space-albinism of the mixed matrices and the joint diagonalization of the time-frequency matrices are simulated to separate the convoluted mixtures. Two kinds of time-frequency analysis methods, Wigner-Ville distribution and improved Wigner-Ville distribution, are introduced, which are calculated with MATLAB 7.0 software. The simulated results show the improved Wigner-Ville distribution method has a better performance for blind separating of non-stationary convoluted and mixed signals.

Sign in / Sign up

Export Citation Format

Share Document