Search

In recent years, computer technologies have been impactful in the design and development of Digital Musical Instruments (DMIs). As music interaction became prominent in the Human-Computer Interaction (HCI) field, emphasis on user requirement upon the design of musical interfaces has also grown since the last decade. Although designing new DMIs is becoming very popular, it is often determined by the designers and often not reflective of users’ needs. In this study, we explored user requirements for the design of a virtual musical instrument of the Malay bonang, an instrument found in the Malay gamelan ensemble. The requirements were elicited from a group of gamelan experts to establish the bonang playing techniques to be mapped to the virtual instrument which we called Air Bonang. Findings revealed that in designing the Air Bonang that is natural and expressive, the fundamental playing techniques of the bonang should be integrated into the system using mid-air interaction. In addition, exploratory techniques might also be integrated into the Air Bonang to leverage musical expression. The outcome of the study proposes design criteria that encompass three aspects of a natural Air Bonang, namely, embodiment, expressiveness, and feedback.

Current road accident research focuses mainly on the key role and importance of Artificial Intelligence (AI) in road accident analysis and prevention. After reviewing the literature, this study found that AI had a wide range of potential applications. It can analyse traffic data more accurately and quickly through advanced machine learning and deep learning technologies, and identify accident risks and dangerous driving behaviours, thereby helping to predict and avoid accidents. Furthermore, the research provides insight into the different types of traffic accidents and the severity of injuries they cause, highlighting the importance of understanding these differences to improve road safety and help inform decision-making. The paper has attempted to develop a comprehensive and diverse road crash impact model by exploring some of the elements that influence road crashes, including human factors, vehicle factors and road environment factors. Finally, this paper identifies some innovations and future research directions for this study, including addressing imbalances and quality issues in collecting and processing data, improving the interpretability and transparency of injury severity expressions, and adopting a more comprehensive approach to analysing road crashes. These innovations will promote greater theoretical and practical progress in road crash research to improve road safety and reduce the damage caused by accidents.

Agricultural images such as fruits and vegetables have previously been recognised and classified using image analysis and computer vision techniques. Mangoes are currently being classified manually, whereby mango sellers must laboriously identify mangoes by hand. This is time-consuming and tedious. In this work, TensorFlow Lite was used as a transfer learning tool. Transfer learning is a fast approach in resolving classification problems effectively using small datasets. This work involves six categories, where four mango types are classified (Harum Manis, Langra, Dasheri and Sindhri), categories for other types of mangoes, and a non-mango category. Each category dataset comprises 100 images, and is split 70/30 between the training and testing set, respectively. This work was undertaken with a mobile-based application that can be used to distinguish various types of mangoes based on the proposed transfer learning method. The results obtained from the conducted experiment show that adopted transfer learning can achieve an accuracy of 95% for mango recognition. A preliminary user acceptance survey was also carried out to investigate the user’s requirements, the effectiveness of the proposed functionalities, and the ease of use of its proposed interfaces, with promising results.

Training and education have become increasingly crucial in obtaining new skills in a variety of fields, especially in assembly and disassembly operations. The main issue in mechanical engineering, particularly in the assembly department, was that automobile engine components assembly was found to be complicated and challenging to assemble using an existing method, where they only rely on a video-based method. The purpose of this paper is to create interactive Virtual Reality (VR) and Augmented Reality (AR) applications that allow users to efficiently assist and complete the assembly tasks. In this work, the authors designed and developed a fully immersive VR application using an HTC Vive headset and two AR applications (marker-less AR application and marker-based AR application) using EPSON MOVERIO BT-300 (AR Smart Glasses). Fourteen engineering students from Universiti Putra Malaysia were selected for the experiment. They were divided into four groups: video-based group, VR-based group, marker-less AR group, and marker-based AR group. They are required to complete all four experiments (video-based experiment, VR-based experiment, marker-less AR experiment, and marker-based AR experiment). The results showed that the marker-less AR application is the best impressive method (37% better), the VR application is the second impressive method (23% better) followed by the marker-based AR application is the third impressive method (3% better) compared to the existing video-based guideline. Therefore, the students favored AR and VR applications rather than the existing method to be used in automobile engine assembly tasks.

The rapid advancement in artificial intelligence (AI) has revolutionized the creation of synthesized multimedia and given rise to DeepFake, a highly realistic fake video or image depicting a person doing or saying something he/ she has never done or said in reality. Attackers use DeepFake to tarnish individuals’ reputations and disseminate fake news, which in turn undermines societies’ stability and security. In response to this cyber security threat, many DeepFake detection methods have been proposed, which show outstanding performance in detecting high-quality DeepFake videos. However, their performance decreases when detecting compressed fake videos. This article investigates the problem of compressed DeepFake video detection. Firstly, popular detection methodologies are reviewed focusing on their abilities to distinguish between real and fake compressed video footage. Then, we attempt to identify and discuss the weaknesses of the methods by examining factors that contribute to decreasing detection efficiency. At the end of this article, we present new generation DeepFake detector techniques that reportedly exhibit improved performance and robustness against video compression. We hope that the contribution from this work inspires innovation for more reliable solutions to combat potential security threats posed by DeepFake videos.

In this century, the use of computer simulation to increase user experience called Virtual Reality (VR) has expanded in the real-life learning contact, which was adopted into practice since 1980s and quick to become a trend in the computer-based simulation world. This study was determined to increase and improve the aspects of trainings, which involves remembering the detailed engine parts assembly among engineering graduates. Currently, the practice requires actual training in the field of automobile engine components assembly. In order to solve this problem, this on-going research proposes a special immersive virtual engine components assembly for students? selfdirected learning. In this study, five main software were used such as CATIA V5 software to produce the 3D engine components, Blender software, Unity3D, Steam VR software and Vive software, and most importantly the HTC vive hardware for VR immersion so that students can appreciate it physically and emotionally through the new virtual interface. Based on the experiment, the data from the respondent will be collected and analysed, and specific test to determine the effectiveness of this application will also be compared with the traditional method.

A grain is a small, hard, dry seed, harvested for human or animal consumption. Almost all the grains have similar shape which are small and round, or small and cylindrical. Not only having similar shape, even they are sometimes similar in colours where mostly consist of brown, yellow and white. Thus, it is hard to differentiate the grains especially among manufacturing companies that handle lots of grains to separate them according to their category. This work aims to contribute to an automatic grain recognition using an image-based query instead of a text-based query. Colour Moment and Wavelet Moment are computed as feature vectors and Support Vector Machine (SVM) algorithm is used to classify the grains based on the extracted features. For evaluation of the proposed prototype, 10-fold cross validation experiment is conducted on five Malaysia’s most used grains which are corn, rice, wheat, barley, and soya. 80% of the images are used for training whereas the remaining 20% images are used for testing. Based on the conducted recognition accuracy testing, it is shown that the feature extraction method mentioned above has successfully obtained an average of 94.7% classification accuracy for grain recognition.

Many applications apply the concept of image recognition to help human in recognising objects simply by just using digital images. A content-based building recognition system could solve the problem of using just text as search input. In this paper, a building recognition system using colour histogram is proposed for recognising buildings in Ipoh city, Perak, Malaysia. The colour features of each building image will be extracted. A feature vector combining the mean, standard deviation, variance, skewness and kurtosis of gray level will be formed to represent each building image. These feature values are later used to train the system using supervised learning algorithm, which is Support Vector Machine (SVM). Lastly, the accuracy of the recognition system is evaluated using 10-fold cross validation. The evaluation results show that the building recognition system is well trained and able to effectively recognise the building images with low misclassification rate.

Object recognition is an approach to identify objects in images or videos. It is the fundamental to voice recognition, handwriting recognition, and also agricultural plant and fruit-based recognitions. There are many works on fruit-based recognition but there is only a limited number for durian species. Durians are quite different physically compared to other fruits. However, all types of durian, be it of different species, have pretty many similar characteristics from one to another. This makes it very hard to differentiate them. Previous works either focused on text-based input or represented the features based on the seam of a durian which may not be suitable to characterise various durian species. There are many ways to differentiate the durian species and this proposed work focused on the curvilinear thorns of the durian as the main approach. Two-dimensional Fast Fourier Transform (2D FFT) with wedge wrapping is applied to the durian images to obtain the Fourier samples where it is then being normalised. Inverse 2D FFT is computed on the normalised Fourier samples to obtain the final Fast Discrete Curvelet Transform features. Through this proposed approach, the curvilinear details of the objects and its edges can be represented better. K-Nearest Neighbour is then trained to effectively classify the durian species. A total of 600 durian images are collected to act as the dataset. 10-fold cross validation is used for evaluation where 80% of the dataset is used for training and the remaining 20% is used for testing. We have conducted several experiments to evaluate various components of the proposed approach. This has resulted to a more effective durian species recognition approach in comparison to the benchmark methods.

Malaysia is well-known for its variety of fruits available in the country such as pineapple, guava, durian, apple, and watermelon. Therefore, it is important for us to get to know more about fruits so that we can take advantage of all the benefits that each fruit can offer. However, problems may arise where a person may know nothing about a particular fruit apart from only having an image of it. Most of the fruit encyclopedias nowadays still rely on text as search input. Furthermore, various features are commonly utilised for representation which can lead to high computational complexity. Therefore, to overcome these problems, a content-based texture-only fruits recognition that accepts an image as input instead of text is proposed. A framework which extracts five texture features (homogeneity, energy, entropy, correlation, and contrast) based on Gray-level Co-occurrence Matrix (GLCM) descriptor is constructed. k-Nearest Neighbour (k-NN) is used at the classifier model to determine the type of fruits. The conducted empirical study has shown that the proposed work has the ability to effectively recognize fruit images with 100% accuracy.

Many fruit recognition approaches today are designed to classify different type of fruits but there is little effort being done for content-based fruit recognition specifically focuses on durian species. Durian, known as the king of tropical fruits, have few similar characteristics between different species where the skin have almost the same colour from green to yellowish brown with just slightly different shape and pattern of thorns. Therefore, it is hard to differentiate them with the current methods. It would be valuable to have an automated content-based recognition framework that can automatically represent and recognise a durian species given a durian image as the input. Therefore, this work aims to contribute to a new representation method based on multiple features for effective durian recognition. Two features based on shape and texture is considered in this work. Simple shape signatures which include area, perimeter, and circularity are used to determine the shape of the fruit durian and its base while the texture of the fruit is constructed based on Local Binary Pattern. We extracted these features from 240 durian images and trained this proposed method using few classifiers. Based on 10-fold cross validation, it is found that Logistic Regression, Gaussian Naïve Bayesian, and Linear Discriminant Analysis classifiers performed equally well with 100% achievement of accuracy, precision, recall, and F1-score. We further tested the proposed algorithm on larger dataset which consisted of 42337 fruit images (64 various categories). Experimental results based on larger and more general dataset have shown that the proposed multiple features trained on Linear Discriminant Analysis classifier able to achieve 72.38% accuracy, 73% precision, 72% recall, and 72% F1-score.

Many fruit recognition systems today are designed to classify different type of fruits but there is no content-based fruit recognition system focuses on durian species. Durian, known as the king of tropical fruits, have few similar characteristics between different species where the skin have almost the same color from green to yellowish brown with slightly different shape of thorns and it is hard to differentiate them with the current methods. Sometimes it is even hard for general consumers to differentiate durian species by themselves. This work aims to contribute to an automatic content-based durian species recognition that will be able to assist users in differentiating various species of durian. Few global contour-based and region-based shape descriptors such as area, perimeter, and circularity are computed as feature vectors and K-Nearest Neighbors algorithm is used to classify the durian based on the extracted features. 10-fold cross-validation is used to evaluate the proposed system. Experimental results have shown that the proposed feature extraction method for the durian species recognition system has successfully obtained a positive recognition rate of 100%.

Identifying the specific machining processes used to produce specific workpiece surfaces is very useful in materials inspection. Machine vision can be used to semi- or fully automate this identification process by firstly extracting features from captured workpiece images, followed by analysis using machine learning algorithms. This enables inspection to be performed more reliably with minimal human intervention. In this paper, three visual texture features are investigated to classify machined workpiece surfaces into the six machining process classes of Turning, Grinding, Horizontal Milling, Vertical Milling, Lapping, and Shaping. These are the multi-directional Gabor filtered images, intensity histogram and edge features statistics. Support Vector Machines (SVM) applying different kernel functions are investigated for best classifier performance. Results indicate that the Gabor-based SVM-linear kernel provides superior performance.

Mitral Annular Displacement (MAD) in echocardiography has been described as a variation in mitral annulus position between the enddiastolic and the end-systolic in a complete cardiac cycle. It could be used as a rapid and reproducible method of determining the LV global systolic function and could be an easily detectable index for wall motion abnormalities. In this study, a computational method of MAD was implemented based on the mitral annulus motion tracking at both sides; namely the lateral side and the septal side using 2D-Echocardiographic (2DE) datasets. This method comprises three main processing stages: 2DE dataset preparation, Region Of Interest (ROI) selection and MAD measurements. For each 2DE dataset, MAD was computed as the movement distance toward the LV apex at both sides individually in twoconsecutive frames using the ‘Euclidian distance' method. Then, the maximum displacement occurs during a complete cardiac cycle was measured in millimetres (mm) for each side. The overall datasets used are 178 original 2D-echocardiography images in 4-chamber view. The experimental results for MAD measurements were compared with results that obtained by TMQ Advanced technique using QLAB software. The comparative analysis was done qualitatively by visual observation of two expert and the comparison scores show that the proposed method of MAD measurements has high acceptability of 85%. Furthermore, the quantitative analysis of the MAD method is comparable with TMAD measurements by QLAB and there is no significant differences in displacement measurements.

Nowadays classifying images into categories have taken a lot of interests in both research and practice. Content Based Image Retrieval (CBIR) was not successful in solving semantic gap problem. Therefore, Bag of Visual Words (BoVW) model was created for quantizing different visual features into words. SIFT detector is invariant and robust to translation, rotations, scaling and partially invariant to affine distortion and illumination changes. The aim of this paper is to investigate the potential usage of BoVW Word model in animal recognition. The better SIFT feature extraction method for pictures of the animal was also specified. The performance evaluation on several SIFT feature strategies validates that MSDSIFT feature extraction will get better results.

In this real world, being able to identify the signs of imminent abnormal behaviors such as aggression or violence and also fights, is of extreme importance in keeping safe those in harm’s way. This research propose an approach to figure out human aggressive movements using Horn-Schunck optical flow algorithm in order to find the flow vector for all video frames. The video frames are collected using digital camera. This research guides and discovers the patterns of body distracted movement so that suspect of aggression can be investigated without body contact. Using the vector of this method, the abnormal and normal video frames are then classified and utilized to define the aggressiveness of humans. Preliminary experiment result showed that the low level of feature extraction can classify human aggressive and non-aggressive movements.

Being able to identify machining processes that produce specific machined surfaces is crucial in modern manufacturing production. Image processing and computer vision technologies have become indispensable tools for automated identification with benefits such as reduction in inspection time and avoidance of human errors due to inconsistency and fatigue. In this paper, the Support Vector Machine (SVM) classifier with various kernels is investigated for the categorization of machined surfaces into the six machining processes of Turning, Grinding, Horizontal Milling, Vertical Milling, Lapping, and Shaping. The effectiveness of the gray-level histogram as the discriminating feature is explored. Experimental results suggest that the SVM with the linear kernel provides superior performance for a dataset consisting of 72 workpiece images.

In automated pulmonary nodules extraction and lung disease diagnosis by image processing techniques, image segmentation is utilized as a primary and the most essential step of lung tumour analysis. But due to extensive similarity between pulmonary vessels, bronchus and arteries in lung region and the low contrast of the Computed-Tomography (CT) image the accuracy of lung tumour diagnosis is highly dependent on the precision of segmentation. Therefore, precise lung CT image segmentation has become a challenging preprocessing task for every lung disease pathological application. In this study, a novel Four-Directional Thresholding (FDT) technique is introduce d. This propounded technique segments the pulmonary parenchyma in Computed-Tomography (CT) images using the Similarity-Based Segmentation (SBS). The proposed technique aims to augment the precision of the CT image thresholding by implementing an advanced thresholding approach from four different directions in which the determination of pixels’ value as being either on foreground or background is highly dependent on its adjacent pixel’s intensity value and the final decision is made based on all four directions’ thresholding results. In this study the importance of neighbour pixels in precision of thresholding with FDT technique is demonstrated and the effectiveness of FDT method has been evaluated on different CT images. Eventually the result of segmentation using FDT method is compared by other precursors techniques, which corroborates the high exactitude of proposed technique.

Local Binary Pattern (LBP) is invariant to the monotonic changes in the grey scale domain. This property enables LBP to present a texture descriptor being useful in applications dealing with the local illumination changes. However, the existing versions of LBP are not able to handle image illumination changes, especially in outdoor environments. The non-patterned illumination changes disturb performance of the background extraction methods. In this paper, an extended version of LBP which is called BackGround LBP (BGLBP) is presented. BGLBP is designed for the background extraction application but it is extendable to the other areas as a texture descriptor. BGLBP is an extension of D-LBP, Centre-Symmetric LBP, ULBP, and R-LBP and it has been designed to inherit the positive properties of previous versions. The performance of BGLBP as a part of background extraction method is investigated. In addition, a comparison between BGLBP as a general texture descriptor and a number of LBP versions is conducted.

Accurate detection of the End-Diastolic (ED) and End-Systolic (ES) frames of a cardiac cycle are significant factors that may affect the accuracy of abnormality assessment of a ventricle. This process is a routine step of the ventricle assessment procedure as most of the time in clinical reports many parameters are measured in these two frames to help in diagnosing and dissection making. According to the previous works the process of detecting the ED and ES remains a challenge in that the ED and ES frames for the cavity are usually determined manually by review of individual image phases of the cavity and/or tracking the tricuspid valve. The proposed algorithm aims to automatically determine the ED and ES frames from the four Dimensional Echocardiographic images (4DE) of the Right Ventricle (RV) from one cardiac cycle. By computing the area of three slices along one cardiac cycle and selecting the maximum area as the ED frame and the minimum area as the ES frame. This method gives an accurate determination for the ED and ES frames, hence avoid the need for time consuming, expert contributions during the process of computing the cavity stroke volume.

In this paper, a technique to classify Engineering Machined Textures (EMT) into the six classes of Turning, Grinding, Horizontal-Milling, Vertical-Milling, Lapping and Shaping, is presented. Multidirectional Gabor features are firstly extracted from each image followed by a dimensionality reduction step using Principal Components Analysis (PCA). The images are finally classified using a supervised Artificial Neural Network (ANN) classifier. Experimental results using a 72-image dataset demonstrate that PCA is able to reduce computational time while improving classification accuracy. In addition, the use of the proposed Gabor filter seems to be more robust compared to other existing techniques.

This paper presents a hybrid method to extract endocardial contour of the right ventricular (RV) in 4-slices from 3D echocardiography dataset. The overall framework comprises four processing phases. In Phase I, the region of interest (ROI) is identified by estimating the cavity boundary. Speckle noise reduction and contrast enhancement were implemented in Phase II as preprocessing tasks. In Phase III, the RV cavity region was segmented by generating intensity threshold which was used for once for all frames. Finally, Phase IV is proposed to extract the RV endocardial contour in a complete cardiac cycle using a combination of shape-based contour detection and improved radial search algorithm. The proposed method was applied to 16 datasets of 3D echocardiography encompassing the RV in long-axis view. The accuracy of experimental results obtained by the proposed method was evaluated qualitatively and quantitatively. It has been done by comparing the segmentation results of RV cavity based on endocardial contour extraction with the ground truth. The comparative analysis results show that the proposed method performs efficiently in all datasets with overall performance of 95% and the root mean square distances (RMSD) measure in terms of mean ± SD was found to be 2.21 ± 0.35 mm for RV endocardial contours.

Text visualization is a well-known technique used in note taking and obviating the difficulties of the learners in focusing on important portions of the text materials to be learnt. On the other hand applying the text visualization methods on the extracted sections required by the learners will amplify the efficiency of perceiving from the text. This paper intends to propose an automatic approach for subjective extracted text visualization to improve leaning from digital text. The main function of this approach divided into two parts, firstly: sentence extraction based on the inclusion of a significant part of speech and role of the words in sentence (i.e. Noun/Names/Places/dates Numbers/ Events) by using the parse tree, secondly: Visualization of the extracted information for maintaining the user’s attention on the content were developed on the basis of psychological, multimedia, and typographic cueing foundations and techniques (visual effects). This experiment conducted on different sample texts in the field of “History” in order to visualize the extracted part.

The present invention relates to a method for automatic motion tracking of ventricular endocardial surface in three dimensional (3D) echocardiography, characterized by the steps of extracting a plurality of ventricular endocardial contours over a complete cardiac cycle; identifying a plurality of landmarks on each ventricular endocardial contour; measuring displacement vector flow (DVF) for each landmark by comparing a pair of consecutive ventricular endocardial contours; measuring velocity vector flow (VVF) for each landmark from end-diastolic (ED) to end-systolic (ES) and vice versa; identifying at least four landmarks from the plurality of landmarks on each ventricular endocardial contour to represent anatomical landmarks of left lateral surface, right lateral surface, inferior wall and anterior wall by using geometrical distance calculation (GDC) algorithm; analysing ventricular endocardial motion direction using a fuzzy logic analyzer (FLA) for the four landmarks identified; updating values of the displacement vector flow (DVF) and velocity vector flow (VVF) based on the ventricular endocardial motion direction; and generating graphical curves of time versus values of the displacement vector flow (DVF) and velocity vector flow (VVF) for the four landmarks identified.