Fuliang Yin

A blind source separation based method for speech encryption

The underdetermined problem poses a significant challenge in blind source separation (BSS) where the number of the source signals is greater than that of the mixed signals. Motivated by the fact that the security of many cryptosystems relies on the apparent intractability of the computational problems such as the integer factorization problem, we exploit the intractability of the underdetermined BSS problem to present a novel BSS-based speech encryption by properly constructing the underdetermined mixing matrix for encryption, and by generating the key signals that satisfy the necessary condition for the proposed method to be unconditionally secure. Both extensive computer simulations and performance analyses results show that the proposed method has high level of security while retaining excellent audio quality

A Region-Growing Permutation Alignment Approach in Frequency-Domain Blind Source Separation of Speech Mixtures

The convolutive blind source separation (BSS) problem can be solved efficiently in the frequency domain, where instantaneous BSS is performed separately in each frequency bin. However, the permutation ambiguity in each frequency bin should be resolved so that the separated frequency components from the same source are grouped together. To solve the permutation problem, this paper presents a new alignment method based on an inter-frequency dependence measure: the powers of separated signals. Bin-wise permutation alignment is applied first across all frequency bins, using the correlation of separated signal powers; then the full frequency band is partitioned into small regions based on the bin-wise permutation alignment result. Finally, region-wise permutation alignment is performed in a region-growing manner. The region-wise permutation correction scheme minimizes the spreading of the misalignment at isolated frequency bins to others, hence to improve permutation alignment. Experiment results in simulated and real environments verify the effectiveness of the proposed method. Analysis demonstrates that the proposed frequency-domain BSS method is computationally efficient.

A blind source separation-based method for multiple images encryption

Blind source separation (BSS) has been successfully applied to many fields such as communications and biomedical engineering. Its application for image encryption, however, remains largely unexplored. In this contribution, a novel BSS-based scheme is proposed for encrypting multiple images, in which the underdetermined BSS problem is fully exploited to achieve the image security. The necessary conditions for generating the key images required for this underdetermined system are presented. The sufficient conditions for constructing the underdetermined mixing matrix for encryption are then described. Extensive computer simulations, coupled with the performance analyses, demonstrate the high level of security of the proposed method.

Blind Source Separation Based on Time-Domain Optimization of a Frequency-Domain Independence Criterion

A new technique for the blind separation of convolutive mixtures is proposed in this paper. Inspired by the works of Amari, Sabala , and Rahbar, we firstly start from the application of Kullback-Leibler divergence in frequency domain, and then we integrate Kullback-Leibler divergence over the whole frequency range of interest to yield a new objective function which turns out to be time-domain variable dependent. In other words, the objective function is derived in frequency domain which can be optimized with respect to time domain variables. The proposed technique has the advantages of frequency domain approaches and is suitable for very long mixing channels, but does not suffer from the local permutation problem as the separation is achieved in time-domain

Blind-Source Separation Based on Decorrelation and Nonstationarity

In this paper, discrete-time blind-source separation (BSS) of instantaneous mixtures is studied. Decorrelation-based sufficient criteria for BSS of stationary and nonstationary sources are derived based on nonstationarity and nonwhiteness. A gradient algorithm is proposed based on these criteria. A batch-data algorithm and an on-line algorithm are developed based on the corollaries of the BSS criteria. These algorithms are especially useful for the separation of nonstationary sources. They are robust to additive white noises if the time-delayed decorrelation and the nonstationarity of the sources are considered simultaneously in the algorithms. Experiment results show the effectiveness and performance of the proposed algorithms

Distributed imm-unscented kalman filter for speaker tracking in microphone array networks

In this paper, we first propose a distributed unscented Kalman filter (DUKF) to overcome the nonlinearity of measurement model in speaker tracking. Next, for the different motion dynamics of a speaker in the in-door environment, we introduce the interacting multiple model (IMM) algorithm and propose a distributed interacting multiple model-unscented Kalman filter (IMM-UKF) for estimating time-varying speakers positions in a microphone array network. In the distributed IMM-UKF based speaker tracking method, the time difference of arrival (TDOA) of the speech signals received by a pair of microphones at each node is estimated by the generalized cross-correlation (GCC) method, then the distributed IMM-UKF is used to track a speaker whose position and speed significantly vary over time in a microphone array network. The proposed method can estimate speakers positions globally in the network and obtain a smoothed trajectory of the speakers movement robustly in noisy and reverberant environments, and it is scalable for speaker tracking. Simulation and real-world experiment results reveal the effectiveness of the proposed speaker tracking method.

Blind source separation for convolutive mixtures based on the joint diagonalization of power spectral density matrices

This paper studies the problem of blind separation of convolutively mixed source signals on the basis of the joint diagonalization (JD) of power spectral density matrices (PSDMs) observed at the output of the separation system. Firstly, a general framework of JD-based blind source separation (BSS) is reviewed and summarized. Special emphasis is put on the separability conditions of sources and mixing system. Secondly, the JD-based BSS is generalized to the separation of convolutive mixtures. The definition of a time and frequency dependent characteristic matrix of sources allows us to state the conditions under which the separation of convolutive mixtures is possible. Lastly, a frequency-domain approach is proposed for convolutive mixture separation. The proposed approach exploits objective functions based on a set of PSDMs. These objective functions are defined in the frequency domain, but are jointly optimized with respect to the time-domain coefficients of the unmixing system. The local permutation ambiguity problems, which are inherent to most frequency-domain approaches, are effectively avoided with the proposed algorithm. Simulation results show that the proposed algorithm is valid for the separation of both simulated and real-word recorded convolutive mixtures.

Speech Segregation Using Constrained ICA

In natural environment, speech often occurs concurrently with acoustic interference. How to effectively extract speech remains a great challenge. This paper describes a novel constrained Independent Component Analysis (ICA) approach, the ICA with reference (ICA-R), to speech segregation. Different from the traditional ICA which recovers simultaneously all the source signals, the ICA-R extracts only some desired source signals from the mixtures of source signals by incorporating some a priori information into the separation process. We show how the ICA-R can be applied to separate a target speech signal from interfering sounds by exploiting a proper reference signal, which is based on the different characteristic between speech signal and its environmental noises, i.e., the speech signal has pitch and its harmonic frequencies whereas the noises usually do not. Results of computer experiments demonstrate the efficiency of the proposed method.

Secure image communication using blind source separation

In this paper, an image encryption method is proposed by using the linear mixing model of blind source separation (BSS). It can simultaneously encrypt multiple images with the same size by mixing them with the same number of statistically independent key images, the size of which is equal to that of the images to be encrypted. Since these multiple images cover mutually through mixing among them while the key images cover them, and there is not any restriction on the key space, the proposed method has high level of security. Computer simulation results show its validity.

An audio watermarking scheme with neural network

An audio watermarking scheme with neural network is presented in this paper. The hiding watermark, which is the combination of the chaotic sequence and the watermark sequence related to original watermark image, is embedded into the DCT coefficients of the original audio signal. Meanwhile, to improve the ability of de-synchronization attack, the synchronous code is embedded into the original audio signal in the time domain. To extract the watermark sequence, first we select DCT coefficients corresponding to the pseudorandom sequence as the training sample, which can be used to train the neural network. Then the DCT coefficients relating to the watermark sequence is taken as the validation sample. Experimental results show that the proposed method has good robustness under general signal manipulations.