Raymond S. T. Lee

Human gait recognition by the fusion of motion and static spatio-temporal templates

In this paper, we propose a gait recognition algorithm that fuses motion and static spatio-temporal templates of sequences of silhouette images, the motion silhouette contour templates (MSCTs) and static silhouette templates (SSTs). MSCTs and SSTs capture the motion and static characteristic of gait. These templates would be computed from the silhouette sequence directly. The performance of the proposed algorithm is evaluated experimentally using the SOTON data set and the USF data set. We compared our proposed algorithm with other research works on these two data sets. Experimental results show that the proposed templates are efficient for human identification in indoor and outdoor environments. The proposed algorithm has a recognition rate of around 85% on the SOTON data set. The recognition rate is around 80% in intrinsic difference group (probes A-C) of USF data set.

A new representation for human gait recognition: motion silhouettes image (MSI)

Recently, gait recognition for human identification has received substantial attention from biometrics researchers. Compared with other biometrics, it is more difficult to disguise. In addition, gait can be captured in a distance by using low-resolution capturing devices. In this paper, we proposed a new representation for human gait recognition which is called Motion Silhouettes Image (MSI). MSI is a grey-level image which embeds the critical spatio-temporal information. Experiments showed that MSI has a high discriminative power for gait recognition. The recognition rate is around 87% in SOTON dataset by using MSI for recognition. The recognition rate is quite promising. In addition, MSI can also reduce the storage size of the dataset. After using MSI, the storage size of SOTON has reduced to 4.2MB.

Tropical cyclone identification and tracking system using integrated neural oscillatory elastic graph matching and hybrid RBF network track mining techniques

In this paper, we present an automatic and integrated neural network-based tropical cyclone (TC) identification and track mining system. The proposed system consists of two main modules: 1) TC pattern identification system using neural oscillatory elastic graph matching model (NOEGM); and 2) TC track mining system using hybrid radial basis function (HRBF) network with time difference and structural learning (TDSL) algorithm.For system evaluation, 120 TC cases appeared in the period between 1985 and 1998 provided by National Oceanic and Atmospheric Administration (NOAA) are being used. In TC pattern recognition from satellite pictures, an overall 98% of correct TC pattern segmentation rate and over 97% of correct classification rate are attained. Moreover, for TC track and intensity mining test, promising result of over 86% is achieved with the application of the hybrid RBF network. Comparing with the bureau numerical TC prediction model (OTCM) used by Guam and the enhanced model (TFS) proposed by Jeng et al., the proposed hybrid RBF has attained an over 30% and 18% improvement in forecast errors.

iJADE Web-miner: an intelligent agent framework for Internet shopping

There is growing interest in using intelligent software agents for a variety of tasks, including navigating and retrieving information from the Internet and from databases, online shopping activities, user authentication, negotiation for resources, and decision making. We propose an integrated framework for information retrieval and information filtering in the context of Internet shopping. We focus on applying agent technology, together with Web mining technology, to automate a series of product search and selection activities. It is based on a multiagent development platform, namely, iJADE (intelligent Java agent development environment), which supports various e-commerce applications. The framework comprises an automatic facial authentication utility and six other modules, namely, customer requirements definition, a requirement-fuzzification scheme, a fuzzy agents-negotiation scheme, a fuzzy product-selection scheme, a product-defuzzification scheme, and a product-evaluation scheme. A series of experiments were carried out and favorable results were produced in executing the framework. From an experimental point of view, we used a database of 1,020 facial images that were obtained under various conditions of facial expression, viewing perspective and size. An overall correct recognition rate of over 85 percent was attained. For the product selection test of our fuzzy shopper system, an average matching rate of more than 81 percent was achieved.

An oscillatory elastic graph matching model for recognition of offline handwritten Chinese characters

Proposes a new application of evolutionary computing - the neural oscillatory elastic graph matching model (NOEGM) for the recognition of offline handwritten Chinese characters. NOEGM consists of three main modules, namely: (1) a feature extraction module using a Gabor filter; (2) a character segmentation module using a neural oscillatory model; and (3) a character recognition module using an elastic graph dynamic link model (EGDLM). In order to optimize the networks performance, a genetic algorithm optimization scheme is integrated into the proposed model. In our research, we applied a sample set of 3,000 handwritten Chinese characters and a test set of 1,000 scanned handwritten Chinese documents to a series of invariant tests, including translation, rotation, dilation and distortion. Experimental results reveal that the overall performance of NOEGM has achieved an average correct recognition rate of over 90%.

iJADE FreeWalker – An Intelligent Ontology Agent-based Tourist Guiding System

Recently, Semantic Web has received substantial attention from the research community. Semantic web aims to provide a new framework that can enable knowledge sharing and resuing. Semantic Web is a collection of web technologies that include a number of markup languages such as RDF, OWL and RDFS. These markup languages are used for modeling a domain ontology. Ontology is defined as “a formal, explicit specification of a shared conceptualization”. By using ontology to model resources, humans and computers (software agents) can have a consensus on the resource structure. The use of these technologies allows the creation of a more effective web search system. In this book chapter, we modeled travel domain ontology by using Web Ontology Language (OWL). Instead of inviting an expert to model the ontology, we created the travel ontology by collecting and analyzing the structural information from a number of travel related websites. Besides, we implemented a tourist context-aware guiding system, iJADE FreeWalker, which is constructed by using Semantic Web technologies. iJADE FreeWalker integrates GPS, ontology and agent technologies to support location awareness for providing the precise navigation and classify the tourist information for the users. The system was tested on 30 novice users. 83% of the users felt that the system can help tourists find tourist information in Hong Kong. www.springerlink.com

Invariant character recognition in Dynamic Link Architecture

A model based on the application of the Dynamic Link Architecture (DLA) is presented for the off-line recognition of Chinese characters. It is a revised DLA model employing 4-vector dynamic link assignment instead of the original 3-vector links. A sample set of Chinese characters was used to test the performance of the model under various transformations including translation, reflection, rotation, dilation and distortion. Challenging results are obtained. An improvement of 40% in recognition rate was attained by using the revised DLA model for Chinese character recognition. On the other hand, for the testing of invariant properties, an overall correct recognition rate of 85% was obtained under various transformations.

Face recognition: elastic relation encoding and structural matching

Face recognition relies heavily on feature extraction and the classification of features in the process of pattern recognition. Existing methods tend to address the problem with some tradeoff between the speed and accuracy in the process. In this paper, a system known as elastic graph dynamic link model (EGDLM) is proposed to provide an effective and reliable solution. The model simplifies the traditional dynamic link model and integrates it with the active contour model for feature extraction. The complex facial pattern matching process is reduced to an elastic graph system matching of facial contours. A database of 1020 facial images was used for model testing and experimental results indicate an improvement of average recognition speed by more than 1000 times, and an overall recognition rate of over 85%.

Using LIDAR doppler velocity data and chaotic oscillatory-based neural network for the forecast of meso-scale wind field

Current research based on various approaches including the use of numerical prediction models, statistical models and machine learning models have provided some encouraging results in the area of long-term weather forecasting. But at the level of meso-scale and even micro-scale severe weather phenomena (involving very short-term chaotic perturbations) such as turbulence and wind shear phenomena, these approaches have not been so successful. This paper focuses on the use of chaotic oscillatory-based neural networks for the study of a meso-scale weather phenomenon, namely, wind shear, a challenging and complex meteorological phenomena which has a vital impact on aviation safety. Using LIDAR data collected at the Hong Kong International Airport via the Hong Kong Observatory, we are able to forecast the Doppler velocities with reasonable accuracy and validate our prediction model. Preliminary results are promising and provide room for further research into its potential for application in aviation forecasting.

Human Identification by Using the Motion and Static Characteristic of Gait

In this paper, we propose a gait recognition algorithm that fuses motion and static spatio-temporal templates of sequences of silhouette images, the motion silhouette contour templates (MSCTs) and static silhouette templates (SSTs). The performance of the proposed algorithm is evaluated experimentally using the SOTON dataset and the USF dataset. The proposed algorithm has a recognition rate of around 80% in intrinsic difference group (Probe A-C) of USF dataset which is 14% higher than the baseline algorithm