Mohamad Hanif Md Saad

Review of Person Re-identification Techniques

Person re-identification across different surveillance cameras with disjoint fields of view has become one of the most interesting and challenging subjects in the area of intelligent video surveillance. Although several methods have been developed and proposed, certain limitations and unresolved issues remain. In all of the existing re-identification approaches, feature vectors are extracted from segmented still images or video frames. Different similarity or dissimilarity measures have been applied to these vectors. Some methods have used simple constant metrics, whereas others have utilised models to obtain optimised metrics. Some have created models based on local colour or texture information, and others have built models based on the gait of people. In general, the main objective of all these approaches is to achieve a higher-accuracy rate and lowercomputational costs. This study summarises several developments in recent literature and discusses the various available methods used in person re-identification. Specifically, their advantages and disadvantages are mentioned and compared.

Design and development of intelligent anomalous behaviour and event detection system

This paper discusses the preliminary design and development of an online anomalous behaviour and event detection system. Live video feeds from CCTV camera were obtained and processed. Important biomechanics features were then extracted and analyzed to discriminate and determine the dynamic objects in the current video scene. Human and nonhuman dynamic objects were then stored in the temporal memory of the system and tracked within the video scene. The tracked humans spatio-temporal activities were then tracked to establish their behaviour and subsequently the occurred event. Augmented with ambient information and a-priori normal condition rules, the system will then search for any anomalous behaviour and event from the set of occurring behaviour and events. The preliminary implementation and result of the system, the Intelligent Video Surveillance System (InViSS©) are presented in this paper.

Appearance-based methods in re-identification: A brief review

In this paper a brief review of appearance-based methods of re-identification in surveillance systems is presented. Previous work involving various parts of re-identification task is discussed and described. The paper concludes in term of re-identification challenges and future work direction.

Review on the application of driving simulators for research and training activities in Malaysia

Driving simulators (DS) are extensively used worldwide as tools in research and training related to driving behaviour and road safety. However, DS have yet to be used extensively in Malaysia. While there is no guarantee that the use of DS may have direct positive impact on road safety, it offers an objective and insightful opportunity to measure and understand driving behaviour. This is especially relevant in Malaysia as it struggles to reduce the high number of fatalities due to road crashes. This paper reviews the role of simulators in research and training worldwide to better understand the driving behaviour and how Malaysia can benefit from useful tool.

Event description from video stream for anomalous human activity and behaviour detection

Event description is a way to objectively describe an event and activity in a video stream in a measurable manner such that it can be processed by computer systems. Event Description Language (EDL) is an event description framework which is used to store and describe objective meaningful event and activity occurring in a video stream. This paper discusses the manual application of one such EDL, the InVISS Event Description Language in annotating actions and events in a video stream. The annotated sequence could later be provided to an activity recognition module for automated recognition of activity in a video stream.

IPQDA: a software tool for intelligent analysis of power quality disturbances

This paper presents the Intelligent Power Quality Disturbance Analysis (IPQDA) software tool that is designed for an automatic analysis of power quality (PQ) disturbance. The main capabilities of the software include analysis of disturbance waveforms, identification of a particular type of disturbance and notification of a disturbance. Another important feature of the program is that it can automatically send email or short messaging notifications upon identification of a disturbance to alert the system operator of a disturbance.

Design and development of an indoor avatar robot

The use of robot as a substitute to human biological body in various type of task such as space exploration, inspection, meetings, and surveillance activities are becoming popular nowadays. Robotic avatar, or avatar robot, is a concept in which a robot replaces the real human. This paper reports the design and development of an avatar robot with telepresence function, which we called AvaRobo. The AvaRobo was derived primarily from our previously developed autonomous ground vehicle (AGV) type surveillance robot platform, named JagaBot™. The JagaBot™ was enriched with specialized two-way remote communication and autonomous navigation capabilities for AvaRobo application. Remote users can manipulate and control AvaRobo and use it to inspect and assess the remote environment and interact with other humans at the remote area. The AvaRobo can move autonomously via line tracking technique or manually by executing the commands given by the remote operator.

Suspicious Loitering Detection from Annotated CCTV Feed Using CEP Based Approach

Smart Surveillance System is a critical system that enables automated detection of anomalous activities from live CCTV feed. The main challenge that needs to be addressed by the Smart Surveillance System is the ability to understand and detect the activities that are currently occurring within the CCTV feed. Suspicious loitering is considered one of the anomalous activities that precede unwanted events, such as break-ins, burglary, and robbery. In this research, the Complex Event Processing (CEP) approach was selected as the system development approach for developing a Smart Surveillance System. Four types of similarity search-based event detectors, namely the Multi-Layered Event Detector for General Application (MEGA), Temporally Constrained Template Match Detector (TCD), Sliding Window Detector (SWD), and Weighted Sliding Window Detector (WSWD) were tested and evaluated to determine the best suspicious loitering event detector to be used in the Smart Surveillance System. The input data to the detectors comprised manually annotated real CCTV feed which was subjected to three noise conditions: (i) no-noise (0% noise) annotation, (ii) 25% noisy annotation and (iii) 46.8% noisy annotation. The 46.8% noisy annotation is assumed to reflect the real ambient operating condition of the Smart Surveillance System; while the no-noise condition was assumed to reflect the perfect CCTV feed acquisition and annotation process. The performance of the detectors was measured in terms of sensitivity, specificity, detection accuracy, and the area under the Receiver’s Operating Curve (ROC). The results obtained showed that MEGA is the best overall detector for suspicious loitering detection in ambient operating conditions with detection accuracy of 97.20% and area under ROC curve of 0.6117.

PROBABILITY ANALYSIS IN DETERMINING THE BEHAVIOUR OF VARIABLE AMPLITUDE STRAIN SIGNAL BASED ON EXTRACTION SEGMENTS

THIS PAPER FOCUSES ON ANALYSIS IN DETERMINING THE BEHAVIOUR OF VARIABLE AMPLITUDE STRAIN SIGNALS BASED ON EXTRACTION OF SEGMENTS. THE CONSTANT LOADING (CAL), THAT WAS USED IN THE LABORATORY TESTS, WAS DESIGNED ACCORDING TO THE VARIABLE AMPLITUDE LOADING (VAL) FROM SOCIETY OF ENGINEER (SAE). THE SAE STRAIN SIGNAL WAS THEN EDITED TO OBTAIN THOSE SEGMENTS THAT CAUSE FATIGUE DAMAGE TO COMPONENTS. THE SEGMENTS WERE THEN SORTED ACCORDING TO THEIR AMPLITUDE AND WERE USED AS A REFERENCE IN THE DESIGN OF THE CAL LOADING FOR THE LABORATORY TESTS. THE STRAIN SIGNALS THAT WERE OBTAINED FROM THE LABORATORY TESTS WERE THEN ANALYSED USING FATIGUE LIFE PREDICTION APPROACH AND STATISTICS, I.E. WEIBULL DISTRIBUTION ANALYSIS. BASED ON THE PLOTS OF THE PROBABILITY DENSITY FUNCTION (PDF), CUMULATIVE DISTRIBUTION FUNCTION (CDF) AND THE PROBABILITY OF FAILURE IN THE WEIBULL DISTRIBUTION ANALYSIS, IT WAS SHOWN THAT MORE THAN 70% FAILURE OCCURRED WHEN THE NUMBER OF CYCLES APPROACHED 1.0 X 1011. THEREFORE, THE WEIBULL DISTRIBUTION ANALYSIS CAN BE USED AS AN ALTERNATIVE TO PREDICT THE FAILURE PROBABILITY.

Complex event detection in an intelligent surveillance system using CAISER platform

Interest about security and asset safety escalates due to the increasing crimes in this century. However, almost all existing surveillance systems have limited self-learning ability that only allow real time monitoring and are unable to identify the actual events that take place in the monitored ambient. As such, this research aims to implement a smart surveillance system with embedded Complex Event Processing (CEP) technology to assist the intrusion detection by correlating raw data extracted from different sources. Four classifiers are used in the CEP engine to predict the event occurrences from the raw data sequence pattern acquired from the door sensors and surveillance camera via intelligent rule template matching algorithm. Confusion matrix in terms of sensitivity, specificity and average detection accuracy as well as ROC plot are employed in classifier performance evaluation to quantify the efficiency of the surveillance system developed.