Monday, January 20, 2020

 
These articles have been peer-reviewed and accepted for publication in JICT, but are pending final changes, are not yet published and may not appear here in their final order of publication until they are assigned to issues. Therefore, the content conforms to our standards but the presentation (e.g. typesetting and proof-reading) is not necessarily up to the JICT standard. Additionally, titles, authors, abstracts and keywords may change before publication.
 
HUMAN ACTIVITY DETECTION AND ACTION RECOGNITION IN VIDEOS USING CONVOLUTIONAL NEURAL NETWORKS
Jagadeesh B & Chandrashekar M Patil
Department of Electronics and Communication Engineering, Vidyavardhaka College of Engineering, India
jagadeesh.b@vvce.ac.in; patilcm@vvce.ac.in
 
 
ABSTRACT 
 
Human activity recognition from video scenes has become a significant area of research in the field of computer vision applications. Action recognition is one of the most challenging problems in the area of video analysis and finds applications in human-computer interaction, anomalous activity detection, crowd monitoring and patient monitoring, Several approaches have been presented for human activity recognition using machine learning technique. The main aim of the work is to detect and track of human, and classification of actions for two publicly available video database. In this work, a novel approach of feature extraction from video sequence by combining Scale Invariant Feature Transform (SIFT) and optical flow computation is used where shape, gradient and orientation features are also incorporated for robust feature formulation. Tracking of the human in the video is implemented using Gaussian Mixture Model. Convolutional Neural Network based classification approach is used for database training and testing purpose. The activity recognition performance is evaluated for two public datasets namely Weizmann dataset and KTH dataset with action recognition accuracy of 98.43% and 94.96% respectively. Experimental and comparative study shows that proposed approach outperforms when compared with state-of-art techniques.
 
Keywords: Action recognition, Convolutional Neural Network, Gaussian Mixture Model, Optical flow, SIFT feature extraction.

 
A COMPUTATIONAL MODEL OF TEMPORAL DYNAMICS FOR ANXIETY IN INTERVIEWEE MENTAL STATE
1Naseer Sanni Ajoge, 2Azizi Ab Aziz & 2Shahrul Azmi Mohd Yusof
1Department of Computer Science, Kaduna Polytechnic, Nigeria
2School of Computing, Universiti Utara Malaysia, Malaysia 
ajogenass@yahoo.com; aziziaziz@uum.edu.my; shahrulazmi@uum.edu.my
 
 
ABSTRACT
 
Anxiety is an aversive motivational state that occurs when an individual perceives threat at events. This condition creates harmful effects for candidates during the interview session. An interviewee overwhelmed in such states deploys worry as resources to cope with the threat hence losing the ability to present self positively for favourable assessments. Most of the digital approaches to assist interviewee in this condition are focused on coaching for verbal and non-verbal cues. The aspect of understanding interviewees’ psychological complexities that influence their behavioural tendencies is lacking in these approaches. As the first step in building an intelligent digital based therapy platform to overcome this issue, this article provides a building block to understanding the interviewee anxiety state by means of a computational model. The model is developed based on the conceptual model derived from generalized anxiety disorder theories. The formal model is evaluated using mathematical analysis to determine possible equilibria state and the simulation results are tested against known cases in the literature. The simulation results show that the degree of threat perceived at events is based on task demand and the resources to cope. Threat is the building block of anxiety through worry which is controlled by one’s personality and inherent trait anxiety. The results conform to established facts in literature. Consequently, this model can serve as the basis to build an integrated interviewee mental state model embedded with self-efficacy and motivation constructs as a holistic approach to support interviewee during simulated training and coaching environment.
 
Keywords: Anxiety States, Computational Cognitive Modeling, Intelligent Support Agent, Interviewee Behavioural States
 
MOTION LEARNING USING SPATIO-TEMPORAL NEURAL NETWORK
1Nooraini Yusoff, 1Farzana Kabir-Ahmad, 1Fadzilah Siraj & 2Mohamad-Farif Jemili
1School of Computing, Universiti Utara Malaysia, Malaysia
2Department of Information Technology and Communication, Sultan Abdul Halim Mu'adzam Shah Polytechnic, Malaysia
farzana58, fad173@uum.edu.my; mfarif@polimas.edu.my
 
  
ABSTRACT
 
Motion trajectory prediction is one of the key areas in behaviour and surveillance studies. Many related successful applications have been reported in the literatures. However, most of the studies are based on sigmoidal neural networks in which they neglect some dynamic properties of the data due to the absence of spatio-temporal encoding functionalities. Even though some sequential (motion) learning studies have been proposed using spatio-temporal neural networks, as in those sigmoidal neural networks, the approach used is mainly supervised learning. In such learning, it requires a target signal, in which this is not always available in some applications. For this study, a motion learning using spatio-temporal neural network is proposed. The learning is based on reward-modulated spike-dependent plasticity (STDP), whereby the learning weight adjustment provided by the standard STDP is modulated by the reinforcement. The implementation of reinforcement approach for motion trajectory can be regarded as the major contribution of this study. In ours, learning is implemented through a reward based basis without the need for learning targets. The algorithm has shown good potential in learning motion trajectory particularly in a noisy and dynamic setting. Furthermore, the learning uses a generic neural network architecture, thus learning is adaptable for many applications.
 
Keywords: Motion learning, reinforcement learning, reward-modulated spike dependent plasticity, spatio-temporal neural network.
 
HYBRID CRYPTOGRAPHIC APPROACH FOR INTERNET OF THINGS APPLICATIONS: A LITERATURE REVIEW
Nur Nabila Mohamed, Yusnani Mohd Yussoff, Mohammed A. Saleh & Habibah Hashim
Faculty of Electrical Engineering, Universiti Teknologi MARA, Malaysia
nurnabilamohamed@gmail.com; yusna233@salam.uitm.edu.my; mohamedswm@yahoo.com; habib350@salam.uitm.edu.my 
 
 
ABSTRACT
 
Cryptography is described as the study of encrypting or secret writing of data using logical and mathematical principles to protect information. This technique has grown in importance in computing technologies for banking service, medical system, transportation and other Internet of Things (IoT)-based applications which have been subjected to increasing security concerns. In cryptography, each of the schemes is built with its own strength respectively, but implementation of single cryptographic scheme into the system has some disadvantages. For instance, symmetric encryption method provides a cost-effective technique of securing data without compromising security, however, sharing the secret key is a vital problem. On the other hand, asymmetric scheme solves the secret key distribution issue yet the standalone technique is slow and consumes more computer resources compared to the symmetric encryption. In contrast, hashing function generates a unique and fixed-length signature for a message to provide data integrity but the method is only a one-way function which is infeasible to invert. As an alternative to solve the security weakness of every single scheme, integration of several cryptographic schemes which also called the hybridization technique is being proposed offering the efficiency of securing data and solving the issue of key distribution. Herein, a review study of articles related to hybrid cryptographic approach from the year 2013 until 2018 is presented. Current IoT domains that implemented hybrid approaches have been identified and the review has been done according to the category of the domain. The significant finding from this literature review is the exploration of various IoT domains that implemented hybrid cryptographic techniques for improving the performance in related works. From the findings, it can be concluded that the hybrid cryptographic approach has been implemented in many IoT cloud computing services. Additionally, AES and ECC are found to be the most popular methods used in the hybrid approach due to its computing speed and security resistance among other schemes.
 

Keywords: hybrid cryptographic approach; internet of things; symmetric encryption; asymmetric encryption; cryptographic hash function.


 
SELF-ADAPTIVE MODEL BASED ON GOAL-ORIENTED REQUIREMENTS ENGINEERING FOR HANDLING SERVICE VARIABILITY
Aradea, Iping Supriana and Kridanto Surendro
1School of Electrical Engineering and Informatics, Bandung Institute of Technology, Indonesia
aradea@unsil.ac.id; iping, endro@informatika.org
 
ABSTRACT
 
Service system is currently facing environmental complexity problems, such as the need of a distributed, heterogeneous, decentralized, and interdependent system which operates dynamically and unpredictably. This condition requires the service system to have an ability to adapt in order to realize sustainable functions. The success of service adaptation is determined by its ability to handle variability at runtime. The purpose of this research is to realize service flexibility through variability modeling, which is an extension of previous work to enrich the adaptability view. The methodology was developed through the MAPE-K control loops approach integrated into the adaptive service (service level) element within the adaptive enterprise service system (AESS) metamodel based on goal-oriented requirements engineering (GORE). Service adaptation scenario was prepared through proactive and reactive adaptation mechanisms. For evaluation, the model was applied to the case of configuration management system. The experimental results show that the model is able to adapt to runtime variability and the growth of the service component items shown by the description of the system scalability. The proposed model has a better alternative design in analyzing variability with a total response that can be applicable in normal operation and overload. It also meets the expected level (level-5: adapting) of the adaptive capability maturity model (ACMM) as a standard for assessment of a service system adaptation.
 
Keywords: Self-adaptive systems, service variability, goal-based, MAPE-K, rule-based systems.
 
GENDER CLASSIFICATION ON SKELETAL REMAINS: EFFICIENCY OF METAHEURISTIC ALGORITHM METHOD AND OPTIMIZED BACK PROPAGATION NEURAL NETWORK
Nurul Liyana Hairuddin, Lizawati Mi Yusuf and Mohd Shahizan Othman
1School of Computing, Universiti Teknologi Malaysia, Malaysia
nurulliyanadin1803@gmail; lizawati@utm.my, shahizan@utm.my
 
 
ABSTRACT
 
In forensic anthropology, gender classification is one of the crucial step involved in developing the biological profiles of the skeleton remains. There are several different parts of skeleton remains and every part contains several features. However, not all features could contribute to gender classification in forensic anthropology. Besides that, another limitation that exists in previous research is the absence of parameter optimization for the classifier. Thus, this paper proposed metaheuristic algorithms such as Particle Swarm Optimization (PSO), Ant Colony Algorithm (ACO) and Harmony Search Algorithm (HSA) based feature selection to identify the most significant features of skeleton remains. Once the set of significant features obtained, the learning rate and momentum of Back Propagation Neural Network (BPNN) were optimized.  This is to obtain good combination of parameters in order to produce a better gender classification. This study used 1,538 data samples from Goldman Osteometric Dataset which consists of femur, humerus and tibia parts. Based on the feature selection result, the Optimized BPNN (OBPNN) outperformed other methods for all datasets. The Ant Colony Algorithm-Optimized Back Propagation Neural Network (ACO-OBPNN) produced the highest accuracy for all parts of the skeleton where for femur is 89.44%, humerus with 88.97% and tibia with 87.52% accuracy. Hence, it can be concluded that optimized parameter is capable of providing a better gender classification performance with the best set of features. Due to good gender classification technique, the implication of study can be seen in forensic anthropology area where the process of developing a biological profile can be completed in a short time and the productivity of anthropologist can be increased.
 
Keywords: Back propagation neural network, metaheuristic algorithms, optimization, gender classification.

Universiti Utara Malaysia Press
 Universiti Utara Malaysia, 06010 UUM Sintok
Kedah Darul Aman, MALAYSIA
Phone: +604-928 4816, Fax : +604-928 4792

Creative Commons License

All articles published in Journal of Information and Communication Technology (JICT) are licensed under a Creative Commons Attribution 4.0 International License.

All Right Reserved. Copyright © 2010, Universiti Utara Malaysia Press