Wen-Hsien Fang

TST-MUSIC for joint DOA-delay estimation

A multiple signal classification (MUSIC)-based approach known as the time-space-time MUSIC (TST-MUSIC) is proposed to jointly estimate the directions of arrival (DOAs) and the propagation delays of a wireless multiray channel. The MUSIC algorithm for the DOA estimation is referred to as the spatial-MUSIC (S-MUSIC) algorithm. On the other hand, the temporal-MUSIC (T-MUSIC), which estimates the propagation delays, is introduced as well. Making use of the space-time characteristics of the multiray channel, the proposed algorithm-in a tree structure-combines the techniques of temporal filtering and of spatial beamforming with three one-dimensional (1-D) MUSIC algorithms, i.e., one S-MUSIC and two T-MUSIC algorithms. The incoming rays are thus grouped, isolated, and estimated. At the same time, the pairing of the estimated DOAs and delays is automatically determined. Furthermore, the proposed approach can resolve the incoming rays with very close DOAs or delays, and the number of antennas required by the TST-MUSIC algorithm can be made less than that of the incoming rays.

FSF MUSIC for Joint DOA and Frequency Estimation and Its Performance Analysis

In this paper, we present a tree-structured frequency-space-frequency (FSF) multiple signal classification (MUSIC)-based algorithm for joint estimation of the directions of arrival (DOAs) and frequencies in wireless communication systems. The proposed approach is a novel twist of parameter estimation and filtering processes, in which two one-dimensional (1-D) frequency (F)- and one 1-D space (S)-MUSIC algorithms are employed-in a tree structure-to estimate the DOAs and frequencies, respectively. In between every other MUSIC algorithm, a temporal filtering process or a spatial beamforming process, implemented by a set of complementary projection matrices, is incorporated to partition the incoming rays to enhance the estimation accuracy, so that the incoming rays can be well resolved even with very close DOAs or frequencies, using the 1-D MUSIC algorithms. Also, with such a tree-structured estimation scheme, the estimated DOAs and frequencies are automatically paired without extra computational overhead. Furthermore, some statistical analyses of the undesired residue signals propagating between the 1-D MUSIC algorithms and the mean square errors of the parameter estimates are derived to provide further insights into the proposed approach. Simulations show that the new approach can provide comparable performance, but with reduced complexity compared with previous works, and that there is a close match between the derived analytic expressions and simulation results

Video anomaly detection and localization using hierarchical feature representation and Gaussian process regression

This paper presents a hierarchical framework for detecting local and global anomalies via hierarchical feature representation and Gaussian process regression. While local anomaly is typically detected as a 3D pattern matching problem, we are more interested in global anomaly that involves multiple normal events interacting in an unusual manner such as car accident. To simultaneously detect local and global anomalies, we formulate the extraction of normal interactions from training video as the problem of efficiently finding the frequent geometric relations of the nearby sparse spatio-temporal interest points. A codebook of interaction templates is then constructed and modeled using Gaussian process regression. A novel inference method for computing the likelihood of an observed interaction is also proposed. As such, our model is robust to slight topological deformations and can handle the noise and data unbalance problems in the training data. Simulations show that our system outperforms the main state-of-the-art methods on this topic and achieves at least 80% detection rates based on three challenging datasets.

Joint Transmit/Receive Antenna Selection in MIMO Systems Based on the Priority-Based Genetic Algorithm

In this letter, a novel genetic algorithm (GA) for joint transmit and receive antenna selection in multiple-input-multiple-output (MIMO) systems is presented. The new GA employs a priority mechanism in the crossover and mutation operations in the evolution process. Computer simulations show that the proposed GA can achieve close performance to the optimum algorithm but with much lower computational complexity.

An efficient unified systolic architecture for the computation of discrete trigonometric transforms

In this paper, a novel unified systolic architecture which can efficiently implement various discrete trigonometric transforms (DXT) including the discrete Fourier transform (DFT), the discrete Hartley transform (DHT), the discrete cosine transform (DCT), and the discrete sine transform (DST) is described. Based on Clenshaws recurrence formula, a set of efficient recurrences for computing the DXT is developed first. Then, the inherent symmetry of the trigonometric functions is further exploited to render a hardware-efficient, systolic structure. For the computation of any N-point DXT of interest, the proposed structure requires only about N/2 multipliers and N adders, thus providing substantial hardware savings compared with previous works. Furthermore, the new scheme can be easily adapted to compute any type of DXT with only minor modification. The complete I/O buffers have been addressed as well which allows for a continuous flow of successive blocks of input data and transformed results in natural order.

A hybrid human fall detection scheme

This paper presents a novel video-based human fall detection system that can detect a human fall in real-time with a high detection rate. This fall detection system is based on an ingenious combination of skeleton feature and human shape variation, which can efficiently distinguish “fall-down” activities from “fall-like” ones. The experimental results indicate that the proposed human fall detection system can achieve a high detection rate and low false alarm rate.

Gaussian Process Regression-Based Video Anomaly Detection and Localization With Hierarchical Feature Representation

This paper presents a hierarchical framework for detecting local and global anomalies via hierarchical feature representation and Gaussian process regression (GPR) which is fully non-parametric and robust to the noisy training data, and supports sparse features. While most research on anomaly detection has focused more on detecting local anomalies, we are more interested in global anomalies that involve multiple normal events interacting in an unusual manner, such as car accidents. To simultaneously detect local and global anomalies, we cast the extraction of normal interactions from the training videos as a problem of finding the frequent geometric relations of the nearby sparse spatio-temporal interest points (STIPs). A codebook of interaction templates is then constructed and modeled using the GPR, based on which a novel inference method for computing the likelihood of an observed interaction is also developed. Thereafter, these local likelihood scores are integrated into globally consistent anomaly masks, from which anomalies can be succinctly identified. To the best of our knowledge, it is the first time GPR is employed to model the relationship of the nearby STIPs for anomaly detection. Simulations based on four widespread datasets show that the new method outperforms the main state-of-the-art methods with lower computational burden.

Abnormal crowd behavior detection and localization using maximum sub-sequence search

This paper presents a novel framework for anomaly event detection and localization in crowded scenes. We propose an anomaly detector that extends the Bayes classifier from multi-class to one-class classification to characterize normal events. We also propose a localization scheme for anomaly localization as a maximum subsequence problem in a video sequence. The maximum subsequence algorithm locates an anomaly event by discovering the optimal collection of successive patches with spatial proximity over time without prior knowledge of the size, start and end of the anomaly event. Our localization scheme can locate multiple occurrences of abnormal events in spite of noise. Experimental results on the well-established UCSD dataset show that the proposed framework significantly outperforms state-of-the-art methods up to 53.55% localization rate. This study concludes that the localization framework plays an important role in abnormal event detection.

A video‐based human fall detection system for smart homes

Abstract In recent years, the global population has begun to age rapidly. Automatic fall detection for senior citizens has become an important issue for smart homes. This paper presents a novel video‐based human fall detection system that can detect a human fall in real‐time with a high detection rate. This fall detection system is based on an ingenious combination of skeleton features and human shape variations, which can efficiently distinguish “fall‐down incidents” from “fall‐like” ones. The experimental results indicate that the proposed human fall detection system can achieve a high detection rate and low false alarm rate.

Soft information assisted space-time multiuser detection for highly loaded CDMA

This letter presents an effective space-time multiuser detector (MUD) with the assistance of soft information in multipath code division multiple access (CDMA) channels. The space-time MUD considered is a simple, separable spatial-temporal filter which consists of a single spatial filter and a single temporal filter. Based on the soft-decision outputs determined in the previous iteration, the soft information is then exchanged in the alternating updates of either the spatial filters or the temporal filters. Furnished simulations show that the proposed scheme can offer substantial performance improvement compared with previous works, especially in highly loaded scenarios.