Li-Ran Shen

A new intrusion detection method based on behavioral model

Intrusion detection has emerged as an important approach to network security. A new method for anomaly intrusion detection is proposed based on linear prediction and Markov chain model. Linear prediction is employed to extract features from system calls sequences of the privileged processes which are used to make up of the character database of those processes, and then the Markov chain model is founded based on those features. The observed behavior of the system is analyzed to infer the probability that the Markov chain model of the norm profile supports the observed behavior. A low probability of support indicates an anomalous behavior that may result from intrusive activities. The experiments show this method is effective and efficient, and can be used in practice to monitor the computer system in real time.

Intrusion detection based on hidden Markov model

The intrusion detection technologies of the network security are researched, and the technologies of pattern recognition are used to intrusion detection. Intrusion detection rely on a wide variety of observable data to distinguish between legitimate and illegitimate activities. Hidden Markov Model (HMM) has been successfully used in speech recognition and some classification areas. Since Anomaly Intrusion Detection can be treated as a classification problem, some basic ideas have been proposed on using HMM to model normal behavior. The experiments have showed that the method based on HMM is effective to detect anomalistic behaviors.

Speech enhancement in short-wave channel based on ICA in empirical mode decomposition domain

It is well known that the non-stationary noise is the most difficult to be removed in speech enhancement. In this paper a novel speech enhancement algorithm based on the empirical mode decomposition (EMD) and then ICA is proposed to suppress the non-stationary noise. The noisy speech is decomposed into components by the EMD and ICA-based vector space, and the components are processed and reconstructed, respectively, by distinguishing between voiced speech and unvoiced speech. There are no requirements of noise whitening and SNR pre-calculating. Experiments show that the proposed method performs well suppressing of the non-stationary noise in short-wave channel for speech enhancement.

Speech stream detection based on higher-order statistics

The aim of speech stream detection is to capture the speech stream whose coming is random. The idea of using Higher Order Statistics (HOS) for speech stream detection is based on exploiting the Gaussian suppression that allows the separation of speech from the noise. HOS have inherent properties that make them well suited when dealing with a mixture of Gaussian and nonGaussian process. In addition, the HOS of speech signals have distinctive features that may be exploited to lead a better estimation and a more accurate discrimination between speech and noise. This paper explores the fourth order cumulants of speech signal and presents a new algorithm for speech stream detection. The considerable experimental results in which data comes from the real recorded on spot, show the method performs well.

Speech Hiding Based on Auditory Wavelet

A novel method to embed secret speech into open speech is proposed. The secret speech is coded into binary parameter bits with Mix-Excitation Linear Prediction (MELP) algorithm, and the bits are used to form hiding information sequence. The open speech is automatically divided into voiced frames and unvoiced frame using auditory wavelet transform. One voice frame, the auditory wavelet transform was used to detect pitch, and the pitch is utilized to the current embedding position in open speech. The information hiding procedure is completed by modifying relevant wavelet coefficients. At the receiver, based on the same pitch detection method, the embedding position is found and the hiding bit is recovered. The secret speech can be received after MELP decoding. The experiments show that the method is strongly robust to many attacks such as compression, filter and so on.

Speech stream detection based on one and half spectrum

The aim of speech stream detection is to capture the speech stream which comes randomly. HOS have inherent properties that make them well suited when dealing with a mixture of Gaussian and non-Gaussian processes. One and half spectrum is a special case of HOS; it not only holds the characters but also saves a lot of computing time. This paper explores the one and half spectrum of speech signal and presents a new algorithm for speech stream detection. The results of a lot of experiments, whose data come from the real recorded on spot, show that the method is performed well.

A new method of robust detection for speech stream

Based on the human auditory perceptual properties, a new method for speech detection was proposed. The new method consists of the classical RASTA-PLP filter in logarithmic power spectrum domain and a differential filter in time domain. In this manner, all kinds of noise can be removed efficiently. Proved by experiments, with mass data of these experiments come from the original real recording of wireless communication on spot, this method can detect speech stream efficiently and the new method outperforms the conventional detection method.

Speech Stream Detection in Short-wave Channel Based on Cepstral of Auditory Wavelet Packets

This paper is aimed at non-stationary characteristic of noise and speech signal in short-wave channel. A novel speech stream detection method based on the cepstral of auditory wavelet packet was proposed. The results of experiments show that this method is a good one for speech stream detection, with nice combat noise performance and adaptation

Speech enhancement in short-wave channel based on empirical mode decomposition

A novel speech enhancement method based on empirical mode decomposition is proposed. The method is a fully data driven approach. Noisy speech signal is decomposed adaptively into oscillatory components called Intrinsic Mode Functions (IMFs) using a process called sifting. The empirical mode decomposition denoising involves thresholding each IMFs. A nonlinear function is introduced for amplitude thresholding. And then reconstructs the estimated speech signal using the processed IMFs. The experimental results show significant improvement in output SNR and quality as compared to recently reported results.

A novel anomaly detection using small training sets

Anomaly detection is an essential component of the protection mechanism against novel attacks.Traditional methods need very large volume of purely training dataset, which is expensive to classify it manually. A new method for anomaly intrusion detection is proposed based on supervised clustering and markov chain model, which is designed to train from a small set of normal data. After short system call sequences are clustered, markov chain is used to learn the relationship among these clusters and classify the normal or abnormal. The observed behavior of the system is analyzed to infer the probability that the markov chain of the norm profile supports the observed behavior. markov information source entropy and condition entropy are used to select parameters. The experiments have showed that the method is effective to detect anomalistic behaviors, and enjoys better generalization ability when a small number of training dataset is used only.