Study of Intrusion Detection Systems

Modern network systems have much trouble in security vulnerabilities such as buffer overflow, bugs in Microsoft Internet, sensor network routing protocol too simple, security flaws of applications, and operating systems. Moreover, wireless devices such as smart phones, personal digital assistants （PDAs）, and sensors have become economically feasible because of technological advances in wireless communication and manufacturing of small and low-cost sensors. There are typologies of vulnerabilities to be exploited in these devices. In order to improve securities, many mechanisms are adopted, including authentication, cryptography, access control, and intrusion detection systems （IDS）. In general, intrusion detection techniques can be categorized into two groups： misuse detection and anomaly detection. The misuse detection systems use patterns of weB-known attacks or weak spots of the systems to identify intrusions. The weakness of misuse detection systems is unable to detect any future （unknown） intrusion until corresponding attack signatures are intruded into the signature database. Anomaly detection methods try to determine whether the deviation is from the established normal usage patterns or not. The critical success of anomaly detection relies on the model of normal behaviors.     

Design and Implementation of a Driver＇s Eye State Recognition Algorithm Based on PERCLOS

In order to improve the general detection accuracy of eye state, this paper puts forward an innovative method for judging human eye state based on PERCLOS. After pretreatment of the eye image, Hough transformation is used for ellipse detection and pupil position. The gray projection variance threshold analysis is then used to help make the final detection. Freeman chain and the Snake model algorithm are used for the corner detection and pre- cise calculation of the height of an open eye. Thus the PER- CLOS value and the eye state can be figured out. The performance of our eye state recognition algorithm is validated by more than 1000 images within product database. The statistics result shows that the fatigue detection accuracy rate can meet the need of usage in complex environment.     

Gaze Tracking Based Regions of Interest Detection Using a Web Camera

Regions of interest （ROIs） detection is one of the key techniques in content-based image retrieval. According to the researches on human vision in biological psychology, the interesting regions or objects can be determined by gaze frequency of human beings. The presented research addresses a novel approach of Regions of interest （ROIs） detection which is on the basis of gaze tracking using an ordinary web camera. This approach utilized an ordinary web camera to capture facial images, in which the eye regions were detected by the methods of fast face detection algorithm based on I-Iaar-like features and adaptive eye template matching. Gaze points were then estimated by the neural network whose net-input vectors consist of face position, center of pupil and glint points. Pupil center was computed based on the centroid of the pupil boundary which was obtained through morphological operations. And the glint points were located through the full-range searching in eye regions, aOIs were obtained by computing the density of gaze points. The experimental results show that the proposed method is objectively effective to measure ROIs and performs well for gaze tracking with an ordinary web camera. This approach is good applicable options for usual users and common applications, especially handicapped.     

Tracking a Screen and Detecting Its Rate of Change in 3-D Video Scenes of Multipurpose Halls

An automatic approach is presented to track a wide screen in a multipurpose hall video scene. Once the screen is located, this system also generates the temporal rate of change by using the edge detection based method. Our approach adopts a scene segmentation algorithm that explores visual features(texture) and depth information to perform efficient screen localization. The cropped region which refers to the wide screen undergoes salient visual cues extraction to retrieve the emphasized changes required in rate-ofchange computation. In addition to video document indexing and retrieval, this work can improve the machine vision capability in the behavior analysis and pattern recognition.     

A Novel Ship Detection Approach for Polarimetric SAR Images Based on a Foreground/Background Separation Framework

In this paper, we present a novel ship detection approach for Polarimetric synthetic aperture radar （PolSAR） images based on a Foreground/background separation （FBS） framework, which exploits the statistical dissimilarity of PoISAR data to separate desired targets from the background clutter. Since the FBS framework takes the spatial relations between pixels into consideration and the separation process exploits the inherent characteristics of PolSAR data, the proposed detector is stable to speckle. And the detection process is filter-free, thus it can preserve the edge and shape information of the extracted targets. Experimental results and comparisons with the standard polarimetric detector show that the proposed method is promising. Factors that affect the performance of the proposed detector are also analyzed and discussed in this paper.     

Microblog Burst Keywords Detection Based on Social Trust and Dynamics Model

Microblog has emerged as a popular medium for providing new sources of information and rapid communications, particularly during burst topics. Burst keywords detection from real-time microblog streams is important for burst topics detection. The exiting algorithms may detect fake burst keywords without taking into account the trustworthiness of the users and human＇s daily timetable. Our work is the first to combine the trustworthiness of the users with burst keywords detection. We propose a novel approach to detect burst keywords based on social trust and dynamics model. We adapt basic notions of dynamics from physics and model keywords bursts as momentum change of the keywords. On the analogy of physical dynamics model, this approach defines mass as the trustworthiness of user and position as the frequency of keywords. We compute each keyword＇s burst value by using Moving average convergence/divergence （MACD） and determine whether it is a burst keyword in a given time window. The experimental results on large-scale Sina microblog dataset show that the proposed approach can avoid detecting fake burst keywords.     

Adaptive Detection of SYN Flooding Attacks at Source-end Networks

An adaptive detection method is proposed to detect SYN flooding attacks at source-end networks. This method can adjust itself to the frequent changes of network conditions. Key features of its design include： （1） creating a detection statistic based on the protocol behavior of TCP SYN-SYN/ACK pairs; （2） forming on-line estimations of the statistical characters of the detection statistic; （3） adjusting its detection threshold according to the variations of network traffic and the latest detection result; （4） decreasing disturbance of random abnormalities in the normal network traffic by consecutive cumulation of threshold violations. Performance analysis and simulation results show the minimum attack traffic that can be detected is about 30% of the legitimate traffic, under the requirements that the probability of false alarms be less than 10^-6, the probability of a miss during an attack be less than 10^-2 and the detection delay be within 7 sampling periods.     

Hybrid System for Robust Faces Detection

The automatic detection of faces is a very important problem. The effectiveness of biometric authentication based on face mainly depends on the method used to locate the face in the image. This paper presents a hybrid system for faces detection in unconstrained cases in which the illumination, pose, occlusion, and size of the face are uncontrolled. To do this, the new method of detection proposed in this paper is based primarily on a technique of automatic learning by using the decision of three neural networks, a technique of energy compaction by using the discrete cosine transform, and a technique of segmentation by the color of human skin. A whole of pictures （faces and no faces） are transformed to vectors of data which will be used for learning the neural networks to separate between the two classes. Discrete cosine transform is used to reduce the dimension of the vectors, to eliminate the redundancies of information, and to store only the useful information in a minimum number of coefficients while the segmentation is used to reduce the space of research in the image. The experimental results have shown that this hybridization of methods will give a very significant improvement of the rate of the recognition, quality of detection, and the time of execution.     

Identifying Periodical Activities Independent of Video Content

In this research, a completely new and accurate method has been presented for detecting periodic activities with the help of machine vision. The proposed method is independent of motion tracking complex algorithms unlike the previous strategies and it is fully independent of contents and types of activities by performing low level calculation. Not using of heavy computations while improving the ability of periodicity detection is regarded as the unique feature of this method. The use of general and flexible framework in this method causes to facilitate the machine vision periodic activities identification process.     

Antarctic Ice-Sheet Freeze-Thaw Detection Based on Improved Physical Model

On the basis of the simple ice-sheet freeze- thaw physical model, a new algorithm of Antarctic ice- sheet freeze-thaw detection was proposed for the automatic threshold segmentation, which did not depend on the priori freeze-thaw distribution. That was the histogram statistics for the data of ice-sheet freeze-thaw physical model by the use of generalized Gaussian model to automatically get the optimal threshold of the dry snow and the wet snow, so as to get the Antarctic freeze-thaw areas. The algorithm improves the computational efficiency, usability and oper- ability of the ice-sheet freeze-thaw detection because the algorithm does not rely on the actual melt information and can automatically select many samples. To some extent, the algorithm also improves the accuracy of the ice-sheet freeze-thaw detection.     

lntra-Cavity Absorption Sensors for Gas Detection Using Wavelength Sweep Technique

Wavelength sweep technique （WST） is introduced into intra-cavity fiber laser （ICFL） for low concentration gas detection. The limitation induced by noise can be eliminated using this method, and the performance of the system is improved. The sensitivity of the system is reduced to less than 300 ppm. With WST, sweeping characteristic of the ICFL can be described according to known gas absorption spectra.     

Jitter Detection for Synchronization Control in Wireless Multimedia Sensor Networks

Collaborative in-network processing operations in Wireless multimedia sensor networks （WMSNs） often require effective synchronization control. Extensive researches in the traditional networks mainly focus on the synchronization control with the buffer management in the receiver. However, for WMSNs, the chaotic transport channel and low bandwidth introduce serious jitter. Jitter degrades the timing relationship among packets in a single media stream and between packets from different media streams and, hence, creates multimedia synchronization problems. Moreover, too much jitter will also degrade the performance of the streaming buffer. By only employing the buffer management scheme in the receiver, we can hardly satisfy the synchronization requirement of the in-network processing. In this study, we propose an active jitter detection mechanism for the synchronization control in WMSNs. This mechanism will improve the quality of service in multimedia networking by discarding the jitter-corrupted packets immediately and balancing the delay and jitter actively. We implement the proposed scheme in the practical WMSNs platform. The experiment results show that our scheme can reduce the average packet jitter effectively and improve the synchronization controlling performance significantly.     

An Ultra Low Steady-State Current Power-on-Reset Circuit in 65nm CMOS Technology

A novel Power-on-reset （POR） circuit is proposed with ultra-low steady-state current consumption. A band=gap voltage eomparator is used to generate a stable pull-up voltage. To eliminate the large current consumptions of the analog part, a power switch is adopted to cut the supply of band-gap voltage comparator, which gained ultra-low current consumption in steady-state after the POR rest process completed. The state of POR circuit is maintained through a state latch circuit. The whole cir- cuit was designed and implemented in 65rim C1V[OS tech- nology with an active area of 120ttm＊160~m. Experimental results show that it has a steady pull=up voltage of 0.69V and a brown-out voltage of 0.49V under a 1.2V supply voltage rising from 0V, plus its steady-state current is only 9hA. The proposed circuit is suitable to be integrated in system on chip to provide a reliable POR signal.     

Labeling Algorithm for Face Detection Using Skin and Hair Characteristics

This research presents an algorithm for face detection based on color images using three main components： skin color characteristics, hair color characteristics, and a decision structure which converts the obtained information from skin and hair regions to labels for identifying the object dependencies and rejecting many of the incorrect decisions. Here we use face color characteristics that have a good resistance against the face rotations and expressions. This algorithm is also capable of being combined with other methods of face recognition in each stage to improve the detection.     

Improving Integrality Objects Based on of Detected Moving Image Matting

The integrality of moving objects is the basis for video-based object tracking and action analysis. But it often becomes unreliable as a result of shadow elimination when the object has similar properties with real shadow. The proposed methods manage to improve the integrality of detected moving objects as much as possible, by image matting operated on candidate shadow regions. Existing approaches for image matting require manual labeling of foreground and background. Considering the moving feature points in shadow may cause classification errors, we propose automatic scribbling methods based on Scale invariant feature transform （SIFT） and Speeded-up robust feature （SURF） respectively. Experiments demon- strate that our methods eliminate real shadow effectively and improve object segmentation in the case of object parts and shadows presenting similar characteristics.     

Multiscale Product for Edge Detection

Many methods have been proposed to extract the most relevant areas of an image. This article explores the method of edge detection by the multiscale product(MP) of the wavelet transform. The wavelet used in this work is the first derivative of a bidimensional Gaussian function. Initially, we construct the wavelet, then we present the MP approach which is applied to binary and grey levels images. This method is compared with other methods without noise and in the presence of noise. The experiment results show that the MP method for edge detection outperforms conventional methods even in noisy environments.     

A Robust Acquisition Scheme for FH Signa in Adverse Environments

Both partial-band jamming and multitone jamming have severe effect on the acquisition of frequency-hopping （FH） signal in adverse environments. In this paper, an anti-jamming FH signal acquisition scheme based on cognitive correlation process is proposed to boost the robustness of acquisition. The main idea of this scheme is to utilize a priori knowledge of FH speed and FH pattern to distinguish jamming signal from received signal. Furthermore, theoretic analysis on detection probability and false probability is given to demonstrate the robust performance of the FH signal acquisition method compared with conventional acquisition scheme without any prior information on FH speed and pattern in adverse environments.     

Real-Time Method for DDoS Attacks Detection Based on Self-Similarity and Wavelet Analysis

Traditional wavelet methods are not efficient to detect the real-time traffic anomaly. To address this drawback, on the basis of self-similarity and wavelet analysis, this paper proposes one real-time method for DDoS attacks detection. Firstly, to effectively analyze new network state, we extracted data packets from network monitor in backward time direction. Secondly, for reducing admissible computing time, we applied time scale selfadjust according to last packets arrival speed. Finally, we adopted three parallel computing strategies to improve real-time performance. Experimental results show that proposed method can quickly and accurately detect DDoS attacks.     

Design and Analysis of Hyperspectral Anomaly Detection Based on Kalman Filter Theory

Anomaly detection generally gains wide attention in hyperspectral imagery for its high spectral resolution. Real-tlme processing is badly needed due to its large data set. This paper presents real-tlme process- ing versions to implement the commonly used RX anomaly detector which make use of information only provided by previous pixels prior to currently being processed pixel. Through these algorithms, hyperspectral image data can be processed timely. Experimental results demonstrate these new real-time versions significantly solve real-time processing problem compared to conventional anomaly de- tector.