Search

Crowd density estimation is a crucial component in surveillance systems to construct safe and efficient urban environments. Due to perspective distortion, individuals in crowd scenes diminish in size as they converge toward the vanishing point. Hence, there are significant visual variations in individuals’ size and appearance, which may lead to inaccurate estimations of crowd counts. This paper proposes an intuitive and effective loss function for the error estimation of crowd counts, particularly in vanishing crowd regions. Specifically, estimation errors in vanishing crowd regions are used to refine and generate network filters that are adaptive toward perspective distortion during network training. Extensive experiments on the challenging UCF-QNRF, WorldExpo, and ShanghaiTech benchmark datasets demonstrate the effectiveness of our novel loss function for training a network to achieve accurate crowd density estimation, particularly in the presence of perspective distortion.

This paper presents a predictive model for fatigue crack growth under loading sequences. Several parameters had been quantified mainly under variable amplitude loading due to its peculiarities. The applicability of the crack closure model was studied to predict the fatigue life span resulting from the continuous effect of loading blocks. The API grade X65 steel, which is widely used in the crude oil and gas industries, was used in this study. Spectrum loading with a two-level block sequences, designated as high-to-low and low-to-high, were designed to correspond to variable amplitude loading. A stress ratio of 0.1 and 0.7 was retained at each block sequences. The effect of block transition from high-to-low caused a delay in crack growth due to the residual plastic wake left behind the crack tip. The crack growth under low-to-high loading blocks were faster than that under high-to-low loading blocks. A modified Walker was established and demonstrated to be capable of describing the features of crack arrest and crack growth acceleration under continuous loading sequences. The normalised root-mean-square error of the predictive model ranged from 0.05 to 0.22 at both load ratios. The integrated crack closure into Walker can reasonably predict the influence of loading sequences on the fatigue crack growth in the API X65 steel.

The paper presents an assessment of fatigue-based reliability in measuring failure random vibration loadings due to unexpected loads under various road conditions. Random loads were the main cause of fatigue failure in a component durability analysis. Acceleration signals were measured during a road test held on rural and highway road surfaces. These signals were captured from an accelerometer placed on a vehicle’s lower arm suspension system in a locally made sedan-type automobile. This arrangement enhanced the capacity for collecting acceleration signals using multi-body dynamics simulation, which would minimise the need for actual acceleration data measurements. Modal analysis was conducted to determine the natural frequencies and mode shapes of the geometry of a coil spring so that strain signals based on each mode shape could be obtained. As modal parameters occurred within the frequency range of road excitation, fatigue life prediction was not affected by instances of dynamic behaviour in the components in time domain. Fatigue life was analysed using the Dirlik, Lalanne and narrowband approaches in frequency domain. The Dirlik method predicted a good correlation for the measured acceleration and generated strain signals at 2.32 × 108, 2.55 × 108 and 1.51 × 108 blocks to failure, indicating that a coil spring could withstand the uneven road surface without failing prematurely. Hence, fatigue-based reliability can evaluate the hazards of rate-reliability, based on predictions of fatigue life data on component durability.

This paper aims to assess the effective strain damage (ESD) model with a durability analysis of an automobile coil spring under consideration of the load sequence effect. Conventional strain-life models such as Coffin-Manson, Morrow and Smith-Watson-Topper are insufficient in predicting fatigue life under the consideration of the load sequence effect. Strain time histories of coil springs were captured based on three types of road conditions: highway, campus and industrial roads. In order to determine the fatigue life, the ESD and conventional strain-life models were employed. The strain-life models predicted lower fatigue life values for campus and industrial roads, which contained many high amplitude events. Similar fatigue life predictions for ESD and conventional strain-life models were found on the highway due to the absence of high amplitude events, which led to a minimal load sequence effect. The suitability of the ESD model was verified through the Weibull probability plot, in which the ESD model fitted well to the distribution compared to conventional strain-life models. The main contribution is that load interactions have a significant influence on predicting fatigue life within the random strain histories of the fatigue properties of a coil spring.

This study aims to assess the fatigue life analysis of strain signals with the implementation of a reconstruction of the signals of the discrete wavelet transform (DWT) decomposition technique. The process involves the time and frequency information used in fatigue life analysis. Strain signals, also known as original signals, were measured based on different road conditions. The DWT method was applied to the original signals to assess the time–frequency domain characteristics. As a result, 14 decomposition signals were obtained and the fatigue life of each signal was evaluated. The decomposition signals contained lower fatigue life, which is below 108 cycles determined at levels 6 to 9. The reconstructed signals were obtained based on these selected levels. The frequency information obtained from the power spectral density (PSD) shows the elimination of high frequency, with low PSD amplitude in the reconstructed signal. The frequency range with high PSD amplitude obtained from this reconstructed signal at 20 Hz was the same as the original signal. Good fatigue life correlation between the reconstructed and original signals enable the prediction of similar fatigue life values, which are at 107 and 105 cycles for the highway and bumpy road conditions, respectively. It can be concluded that the implementation of this technique was able to provide good agreement of strain signals in fatigue life analysis.

Image segmentation is one of the most important tasks in image processing and pattern recognition. One of the most efficient and popular techniques for image segmentation is image thresholding. Among several thresholding methods, Kapur's (maximum entropy (ME)) and Otsu's methods have been widely adopted for their simplicity and effectiveness. Although efficient in the case of bi-level thresholding, they are very computationally expensive when extended to multilevel thresholding because they employ an exhaustive search for the optimal thresholds. In this paper, a fast scheme based on a modified Bees Algorithm (BA) called the Patch-Levy-based Bees Algorithm (PLBA) is adopted to render Kapur's (ME) and Otsu's methods more practical; this is achieved by accelerating the search for the optimal thresholds in multilevel thresholding. The experimental results demonstrate that the proposed PLBA-based thresholding algorithms are able to converge to the optimal multiple thresholds much faster than their corresponding methods based on Basic BA. The experiments also show that the thresholding algorithms based on BA algorithms outperform corresponding state-of-the-art metaheuristic-based methods that employ Bacterial Foraging Optimization (BFO) and quantum mechanism (quantum-inspired algorithms) and perform better than the non-metaheuristic-based Two-Stage Multi-threshold Otsu method (TSMO) in terms of the segmented image quality. In addition, the results show the high degree of stability of the proposed PLBA-based algorithms.

This CAD system calculates proliferation rate estimation (PRE) automatically.A novel Iterative Randomized Irregular Circular Algorithm (IRIC) has been proposed.The Brain Tumor Ki-67 Histology Images are taken from UKM Medical Centre.Prior to a random set region of interest, IRIC counts blue and brown cells and PRE.IRIC outperforms Circular Hough Transform about 98% and of F-measurement rate. Proliferation rate estimation (PRE) is clinically performed from Ki-67 histopathology images. As brain tumor tissues are very complex, accurate PRE determination requires manual cell counting that is tedious, time consuming and inherently inaccurate due to inter-personal variations. Therefore, pathologists usually determine the PRE based on their experience and visualization without actual counting. Automating PRE can substantially increase the efficiency and accuracy of pathologists' determination of PRE. In addition, developing a deterministic and reproducible proliferation rate value is crucial to reduce inter-observer variability. In this paper, a PRE Computer Aided Diagnosis (PRECAD) system has been developed to automate the determination of PRE from Ki-67 histopathology microscopic images for brain tumors. The process involves six steps: color space transformation, customized color modification, nuclei segmentation based on K-Means clustering, preprocessing the extracted cells, counting based on an iterative structured circle detection (IRIC) algorithm, and finally, calculating the PRE value. The proposed IRIC algorithm is able to detect irregular and overlapping cells by introducing dynamic initialization to the basic RCD method, dividing the entire image into partitions based on 8-neighbor connected components. We initiated a new selection method for determining a best circle candidate that yields a reduced probability of incorrectly detecting circles, and proposed a new technique for detecting irregular cells via a dynamic number of iterations that guarantees finding all the cells in a selected partition. Using the same innovations mentioned above, our proposed IRIC algorithm can also be used to detect irregular and ?two or more overlapping cells. The proposed PRECAD system achieved high accuracy, as determined by quantitative analysis of precision, recall and F-measurement test values of 97.8%, 98.3% and 98% for blue cells and 98.7%, 98% and 98.4% for brown cells, respectively. Thus, our proposed PRECAD system is as reliable as a pathologist for estimating the proliferation rate, while also featuring inherent reproducibility.

This study aims to determine the fatigue reliability of an automotive crankshaft by stochastically inducing random loads because of limited experimental data. Given the difficulties in capturing actual load–time history data from laboratory or field-testing and the lack of loading history data, the corresponding stochastic modellings are formulated. The maximum and minimum loads obtained from the automobile industry at various rotational speeds are stochastically induced to generate random loads for assessing fatigue reliability. Fatigue life is evaluated using the Coffin–Manson, Morrow and Smith–Watson–Topper models in the time domain using the rainflow cycle counting technique. Weibull distribution is identified as the best fit of the data for fatigue lives on the basis of Akaike’s information criterion and is used to model reliability and hazard rate. Reliability below 0.09 with a hazard rate of over 1.35 × 10−5 is proposed as the high-risk regions according to the mean number of cycles to failure obtained from the fatigue life models. Thus, stochastic-induced random loads can assess the hazard rate-reliability from fatigue life data to predict the durability and structural integrity of components.

This study aims to assess the risk-based fatigue life from each decomposition signals for various roads through the probabilistic analysis on the suspension system. Cyclic loading in fatigue failure and fatigue life-reliability was the main concern in the design and durability analysis. Different types of road surface signals, namely, rural, urban and highway roads were measured by a road test. These signals were decomposed using the Hilbert-Huang transform that contains the simpler component for describing the information from the original signal based on each decomposition mode. The degree of decomposition can be determined from the IMF level. The decomposition signals were then used in predicting the fatigue life of the given component. Akaike’s information criterion was used to determine suitable distribution of the experimental data. Fatigue life was calculated using the mean-cycle-to-failure (McTF) and the result was used to evaluate risk areas for each road surface. The McTF of rural, urban and highway road surfaces for the Morrow model were obtained at 2.4768 × 104, 1.5585 × 108 and 1.5491 × 106 cycles per block, respectively. The McTF of the Morrow model is the safe region that predicts the longer fatigue life compared to in the Coffin-Manson and SWT model. The risk was predicted to be 1.456 × 10–3 (rural), 1.221 × 10–3 (urban) and 1.890 × 10–3 (highway) failures per cycles per block based on the hazard rate function. Hence, high-risk and low-risk were found to occur at above and below of each risk prediction value of the various roads. Risk-based life cycle assessment from decomposition signals has the capability in assessing the fatigue life and risk area of the component under strain loading.

The capability of structures to absorb large amounts of energy is a crucial factor, particularly for structural components of vehicles, in reducing injury in case of collision. In this study, an experimental investigation was conducted to study the crashworthiness of polymeric foam-filled structures to the pultruded square cross-section E-Glass fiber-reinforced polyester composite tube profiles. Quasi-static compression was applied axially to composite tubes to determine the response of the quasi-static load displacement curve during progressive damage. Three pultruded composite tube wall thicknesses at different sizes were examined, and the effects of crushing behavior and failure modes were analyzed and discussed. Experimental results indicated that the foam-filled profile is superior to the non-filled foam composite tube profile in terms of the capacity to absorb specific energy.

The Bees Algorithm (BA) is a population-based metaheuristic algorithm inspired by the foraging behavior of honeybees. This algorithm has been successfully used as an optimization tool in combinatorial and functional optimization fields. In addition, its behavior very closely mimics the actual behavior that occurs in nature, and it is very simple and easy to implement. However, its convergence speed to the optimal solution still needs further improvement and it also needs a mechanism to obviate getting trapped in local optima. In this paper, a novel initialization algorithm based on the patch concept and Levy flight distribution is proposed to initialize the population of bees in BA. Consequently, we incorporate this initialization procedure into a proposed enhanced BA variant. The proposed variant is more natural than conventional variants of BA. It mimics the patch environment in nature and Levy flight, which is believed to characterize the foraging patterns of bees in nature. The results of experiments conducted on several widely used high-dimensional benchmarks indicate that our proposed enhanced BA variant significantly outperforms other BA variants and state-of-the-art variants of the Artificial Bee Colony (ABC) algorithm in terms of solution quality, convergence speed, and success rate. In addition, the results of experimental analyses conducted indicate that our proposed enhanced BA is very stable, has the ability to deal with differences in search ranges, and rapidly converges without getting stuck in local optima.

This paper presents the investigation of the characteristic behaviour of polymer matrix composites under Charpy impact conditions with different design configurations of the laminate structure. The aim of this study was to evaluate the capability of different lamination designs for composite materials, in term of contact load, energy absorption, deflection and damage behaviour. In this study, laminated panels were fabricated using chopped strand mat (CSM), woven roving fabric (WR) and foam-PVC Klegecell as reinforcement with a combination of epoxy or polyester resin, respectively. Structural panels of composite laminates were produced using a hand lay-up technique. Each configuration design was impact tested to failure. Finite element analyses (FEA) were employed in this study to correlate the experimental value of energy absorption with simulation results. The characteristics of different reinforcement types, matrix type, hybrid type, architecture and orientation type were studied. These characteristics need to be considered, due to their affecting the characteristic behaviour of the composite lamination structures. Based on the results, it was found that differences in configuration design of the lamination structure of the polymer matrix composites do influence the strength and weakness of the materials.

This paper presents a multilevel image thresholding approach which relies on Tsallis entropy using Fuzzy partition with a novel threshold selection technique. In order to compute the optimal threshold values, Differential Evolution (DE) has been employed. The proposed method can further be exploited in image segmentation which is considered to be a critical step in image processing. Our proposed threshold selection technique is based on Tsallis-Fuzzy entropy and the results are compared with Shannon entropy (or fuzzy entropy) and Tsallis entropy based existing threshold selection techniques. The experiments are performed on two different sets of images and the results have been compared with that of existing state-of-the-art methods, namely, Patch Levy Bees' Algorithm (PLBA), Bacterial Foraging optimization (BFO), modified Bacterial Foraging optimization (MBFO) and Bees' Algorithm (BA). Quantitative analysis is carried out based on three image quality metrics viz SSIM, PSNR and SNR. Standard deviation and CPU time for convergence of the objective function have been calculated for performance evaluation. Furthermore, the statistical significance of our method has been estimated using Friedman test and Wilcoxon test. The experimental results manifest that our method produces results superior to the methods in comparison.

Skeletonization and also known as thinning process is an important step in pre-processing phase. Skeletonization is a crucial process for many applications such as OCR, writer identification ect. However, the improvements in this area still remain due to researches recently. A new skeletonization algorithm is proposed in this paper. The algorithm is combining between parallel and sequential which categorized under iterative approach. The proposed method conducted into experiments of benchmark dataset for evaluation. The result is obtaining much better results comparing with other thinning methods is included in comparison part.

This paper proposes an alternative statistical approach that can significantly characterise the fatigue strength property of metallic material. In this study, a simple impulsive excitation technique based on ASTM E1876 is conducted to measure the vibro-acoustic signal simultaneously using data acquisition consisted of an accelerometer-microphone configuration. The considered metallic material in the proposed design is composed of rectangular bar in cross section of medium carbon steel S50C, stainless steel AISI 304, cast iron FCD 500 and brass with a similar size of dimension as a specimen. To take into account the effect of noise existance, an alternative Z-stem filter de-noising technique is derived and applied to minimise the noise frequency components. Moreover, the effectiveness of this filtering technique is calculated. The filtered vibro-acoustic signal has been statistically analysed using an alternative time domain analysis method known as Mesokurtosis Zonal Nonparametric (M-Z-N). The accurate curve fitting equation for every metallic material is obtained using the combination of the calculated M-Z-N coefficient for vibration and acoustic impulse signals. The fatigue strength characterisation is carried out by the comparison of the experimental results with fatigue strength properties obtained in the literature. Correlation between the slopes of the curve fitting and fatigue strength values validate the proposed method.

Robot soccer competition is a new learning method introduced in the Faculty of Information Science and Technology in the past two years. Previously, various friendly matches have been held between the National University of Malaysia team and Malaysian polytechnic students. However, not all of the students, especially the FTSM students felt the fun of the game. Therefore, the competition format is applied in two courses; TC2263 - Knowledge-based Systems and TR3343 – Real Time Systems. Parallel with the purpose of student self-centered learning method, assignments are given to the students to develop the best robot soccer game strategy. They only need to attend some lectures and laboratory sessions and they are given the opportunity to complete the remaining tasks individually until the day of the competition. There are three methodologies used in this research; the data taken during the in class competition where it is judged by external judges who are experts in robot soccer, the observation of the final grade, and finally the questionnaire through the Course Teaching Evaluation System provided by UKM. The application of robot soccer competition in this class has shown increased performance in analytical and critical thinking skills in problem solving, and programming skills and technical report writing. Besides that, the application of robot soccer competition can also realise self-centered learning among FTSM students.

Binary image representation is essential format for document analysis. In general, different available binarization techniques are implemented for different types of binarization problems. The majority of binarization techniques are complex and are compounded from filters and existing operations. However, the few simple thresholding methods available cannot be applied to many binarization problems. In this paper, we propose a local binarization method based on a simple, novel thresholding method with dynamic and flexible windows. The proposed method is tested on selected samples called the DIBCO 2009 benchmark dataset using specialized evaluation techniques for binarization processes. To evaluate the performance of our proposed method, we compared it with the Niblack, Sauvola and NICK methods. The results of the experiments show that the proposed method adapts well to all types of binarization challenges, can deal with higher numbers of binarization problems and boosts the overall performance of the binarization.