Search

Buildings are among the most important elements in the urban structure that can affect urban planning. Therefore, it is important to create the footprint of the buildings, especially in developing cities, which is highly time-consuming and costly. Although LiDAR technology has already been used for this purpose, the need to process voluminous amounts of noisy data and make building footprint extraction in accurate. In this study, we propose a step-by-step analysis of LiDAR data using a rule-based algorithm called DB-creator in order to automatically create building footprints. DB-creator was specifically chose as it does not require external data or region information to construct the footprints. The constructed footprints based from the algorithm was compared with manually created ground truth building footprints to assess accuracy. From experimental results, RMSE for urban and rural areas were ± 0.62 m and ± 0.28 m, respectively, which is highly accurate considering LiDAR’s a 0.5 m surveying distance between two points and 0.6 m distance between rows. Moreover, the kappa coefficient were 0.948 and 0.958 for the urban and rural areas, respectively (which are confirmed by T values of 150.204 and 255.553 at p ≤ 0.01 for the urban and rural areas, respectively). The Standard Errors respectively obtained for urban and rural areas were 0.001 and 0.002, reflecting slight internal variations between the built footprint maps. This also highlights the certainty of the kappa coefficient, indicating that extraction of building footprints is highly accurate.

In recent years complex food security issues caused by climatic changes, limitations in human labour, and increasing production costs require a strategic approach in addressing problems. The emergence of artificial intelligence due to the capability of recent advances in computing architectures could become a new alternative to existing solutions. Deep learning algorithms in computer vision for image classification and object detection can facilitate the agriculture industry, especially in paddy cultivation, to alleviate human efforts in laborious, burdensome, and repetitive tasks. Optimal planting density is a crucial factor for paddy cultivation as it will influence the quality and quantity of production. There have been several studies involving planting density using computer vision and remote sensing approaches. While most of the studies have shown promising results, they have disadvantages and show room for improvement. One of the disadvantages is that the studies aim to detect and count all the paddy seedlings to determine planting density. The defective paddy seedlings’ locations are not pointed out to help farmers during the sowing process. In this work we aimed to explore several deep convolutional neural networks (DCNN) models to determine which one performs the best for defective paddy seedling detection using aerial imagery. Thus, we evaluated the accuracy, robustness, and inference latency of one- and two-stage pretrained object detectors combined with state-of-the-art feature extractors such as EfficientNet, ResNet50, and MobilenetV2 as a backbone. We also investigated the effect of transfer learning with fine-tuning on the performance of the aforementioned pretrained models. Experimental results showed that our proposed methods were capable of detecting the defective paddy rice seedlings with the highest precision and an F1-Score of 0.83 and 0.77, respectively, using a one-stage pretrained object detector called EfficientDet-D1 EficientNet.

This work discuss the task of automatically detecting satire instances in short articles. It is the study of extracting the most optimal features by using a deep learning architecture combined with carefully handcrafted contextual features. It is found that a few sets can perform well when they are used independently, but the others not so much. However, even the latter sets become very useful after the combination process with the former sets. This shows that each of the feature sets are significant. Finally, the combined feature sets undergoes the classification using well-known machine learning classification algorithms. The best algorithm for this task is found to be Logistic Regression. The outcome of all the experiments are good in all the metrics used. The result comparison to existing works in the same domain shows that the proposed method is slightly better with 0.94 in terms of F1-measure, while existing works managed to obtain 0.91 (Yang et al. , 2017), 0.90 (Zhang et al. , 2016), and 0.88 (Rubin et al. , 2016). The performance of each feature sets are also given as additional information. The main purpose of this work is to show that the combination of features extracted using supervised learning with the ones extracted manually can yield a good performance. It is also to open doors for other researchers to take into account the contextual meaning behind a figurative language type such as satire.

Forest conservation is crucial for the maintenance of a healthy and thriving ecosystem. The field of remote sensing (RS) has been integral with the wide adoption of computer vision and sensor technologies for forest land observation. One critical area of interest is the detection of active forest fires. A forest fire, which occurs naturally or manually induced, can quickly sweep through vast amounts of land, leaving behind unfathomable damage and loss of lives. Automatic detection of active forest fires (and burning biomass) is hence an important area to pursue to avoid unwanted catastrophes. Early fire detection can also be useful for decision makers to plan mitigation strategies as well as extinguishing efforts. In this paper, we present a deep learning framework called Fire-Net, that is trained on Landsat-8 imagery for the detection of active fires and burning biomass. Specifically, we fuse the optical (Red, Green, and Blue) and thermal modalities from the images for a more effective representation. In addition, our network leverages the residual convolution and separable convolution blocks, enabling deeper features from coarse datasets to be extracted. Experimental results show an overall accuracy of 97.35%, while also being able to robustly detect small active fires. The imagery for this study is taken from Australian and North American forests regions, the Amazon rainforest, Central Africa and Chernobyl (Ukraine), where forest fires are actively reported.

Building damage maps can be generated from either optical or Light Detection and Ranging (Lidar) datasets. In the wake of a disaster such as an earthquake, a timely and detailed map is a critical reference for disaster teams in order to plan and perform rescue and evacuation missions. Recent studies have shown that, instead of being used individually, optical and Lidar data can potentially be fused to obtain greater detail. In this study, we explore this fusion potential, which incorporates deep learning. The overall framework involves a novel End-to-End convolutional neural network (CNN) that performs building damage detection. Specifically, our building damage detection network (BDD-Net) utilizes three deep feature streams (through a multi-scale residual depth-wise convolution block) that are fused at different levels of the network. This is unlike other fusion networks that only perform fusion at the first and the last levels. The performance of BDD-Net is evaluated under three different phases, using optical and Lidar datasets for the 2010 Haiti Earthquake. The three main phases are: (1) data preprocessing and building footprint extraction based on building vector maps, (2) sample data preparation and data augmentation, and (3) model optimization and building damage map generation. The results of building damage detection in two scenarios show that fusing the optical and Lidar datasets significantly improves building damage map generation, with an overall accuracy (OA) greater than 88%.

The focal point of this work is to automatically detect metaphor instances in short texts. It is the study of extricating the most optimal features for the task by using a deep learning architecture and carefully hand-crafted contextual features. The first feature set is created using a Convolutional Neural Network (CNN) architecture. Then, three other feature sets are manually hand-crafted using contextual justifications. Next, all of the feature sets are combined. Finally, the combined feature sets undergo the classification process using Support Vector Machine, Logistic Regression, Decision Tree, K-Nearest Neighbour and Discriminatory Analysis. These well-known ma-chine learning classification algorithms are used at the same time for the purpose of comparison. The best algorithm for this task is found to be Support Vector Machine (SVM). The outcome of all the experiments using SVM are good in all the metrics used, with F1-measure of 0.83. Finally, comparison to existing works and performance of each feature sets are given. It is also found that a few sets performed well when they are used independently. However, even the sets that are not useful separately is proven to be very useful after the combination process.

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 paper, we investigate the use of Explainable Artificial Intelligence (XAI) methods for the interpretation of two Convolutional Neural Network (CNN) classifiers in the field of remote sensing (RS). Specifically, the SegNet and Unet architectures for RS building information extraction and segmentation are evaluated using a comprehensive array of primary- and layer-attributions XAI methods. The attribution methods are quantitatively evaluated using the sensitivity metric. Based on the visualization of the different XAI methods, Deconvolution and GradCAM results in many of the study areas show reliability. Moreover, these methods are able to accurately interpret both Unet's and SegNet's decisions and managed to analyze and reveal the internal mechanisms in both models (confirmed by the low sensitivity scores). Overall, no single method stood out as the best one.

During or after natural disasters, information about location, cause, and severity, is crucial for early responders to act accordingly. Building damage is one of the major disaster types that occurred repeatedly. Being able to estimate the extent and location of damaged buildings are important so that emergency personnel and rescue teams can expedite efforts to the right building in affected location. Satellite imagery is a powerful visual resource that can be used to assess the extent of damages within a wide geographical area. However, current post-disaster practice requires manual annotation of damaged buildings, which is labor intensive and time consuming. Resultantly, traditional damage detection methods have been outperformed in terms of accuracy by Deep Learning (DL) architectures such as the Convolutional Neural Networks (CNN). Therefore, we developed a novel framework named Multi-scale Siamese Building Damage Assessment Network (MSBDA-Net). The proposed framework includes a two-step approach. The first stage is building localization, which a mask of all buildings before disaster will be generated. The second stage is a multi-scale Siamese damage assessment model, where the network takes the image pairs contained pre- and post-disaster as input and classify building on different damage levels. The evaluation results of proposed method indicate the applicability of the proposed method in both building segmentation (Fl-score=86.3%) and damage assessment (Fl-score=78.44 %).

Tremendous number of f food images in the social media services can be exploited by using food recognition for healthcare benefits and food industry marketing. The main challenges in food recognition are the large variability of food appearance that often generates a highly diverse and ambiguous descriptions of local feature. Ironically, the ambiguous descriptions of local feature have triggered information loss in visual dictionary constructions from the hard assignment practices. The current method based on hard assignment and Fisher vector approach to construct visual dictionary have unexpectedly cause errors from the uncertainty problem during visual word assignation. This research proposes a method of combination in soft assignment technique by using fuzzy encoding approach and maximum pooling technique to aggregate the features to produce a highly discriminative and robust visual dictionary across various local features and machine learning classifiers. The local features by using MSER detector with SURF descriptor was encoded by using fuzzy encoding approach. Support vector machine (SVM) with linear kernel was employed to evaluate the effect of fuzzy encoding. The results of the experiments have demonstrated a noteworthy classification performance of fuzzy encoding approach compared to the traditional approach based on hard assignment and Fisher vector technique. The effects of uncertainty and plausibility were minimized along with more discriminative and compact visual dictionary representation.

This paper proposes a UAV-based PM2.5 air quality and temperature-humidity monitoring system. The system includes an air quality detector comprising four Arduino sensor modules. Specifically, it includes a dust (DSM501A) sensor and a temperature and humidity (DHT11) sensor. The NEO-6M GPS module and DS3231 real-time module are also included for input visualization. A DIY SD card logging shield and memory module is also available for data recording purposes. The Arduino-based board houses multiple sensors and all are programmable using the Arduino integrated development environment (IDE) coding tool. Measurements conducted in a vertical flight path show promise where comparisons with ground truth references data showed good similarity. Overall, the results point to the idea that a light-weight and portable system can be used for accurate and reliable remote sensing data collection (in this case, PM2.5 concentration data and environmental data).

Our work focuses on detecting sarcasm in tweets using deep learning extracted features combined with contextual handcrafted features. A feature set is extracted from a Convolutional Neural Network (CNN) architecture before it is combined with carefully handcrafted feature sets. These handcrafted feature sets are created based on their respective contextual explanations. Each feature sets are specifically designed for the sole task of sarcasm detection. The objective is to find the most optimal features. Some sets are good to go even when it is used in independence. Other sets are not significant without any combination. The results of the experiments are positive in terms of Accuracy, Precision, Recall and F1-measure. The combination of features is classified using a few machine learning techniques for comparison purposes. Logistic Regression is found to be the best classification algorithm for this task. Furthermore, result comparison to recent works and the performance of each feature set are also shown as additional information..

The visual analysis of foods on social media by using food recognition algorithm provides valuable insight from the health, cultural and marketing. Food recognition offers a means to automatically recognise foods as well the useful information such as calories and nutritional estimation by using image processing and machine learning technique. The interest points in food image can be detected effectively by using Maximally Stable Extremal Region (MSER). As MSER used global segmentation and many food images have a complex background, there are numerous irrelevant interest points are detected. These interest points are considered as noises that lead to computation burden in the overall recognition process. Therefore, this research proposes an Extremal Region Selection (ERS) algorithm to improve MSER detection by reducing the number of irrelevant extremal regions by using unsupervised learning based on the k-means algorithm. The performance of ERS algorithm is evaluated based on the classification performance metrics by using classification rate (CR), error rate (ERT), precision (Prec.) and recall (rec.) as well as the number of extremal regions produced by ERS. UECFOOD-100 and UNICT-FD1200 are the two food datasets used to benchmark the proposed algorithm. The results of this research have found that the ERS algorithm by using optimum parameters and thresholds, be able to reduce the number of extremal regions with sustained classification performance.

In recent years, remote-sensing (RS) technologies have been used together with image processing and traditional techniques in various disaster-related works. Among these is detecting building damage from orthophoto imagery that was inflicted by earthquakes. Automatic and visual techniques are considered as typical methods to produce building damage maps using RS images. The visual technique, however, is time-consuming due to manual sampling. The automatic method is able to detect the damaged building by extracting the defect features. However, various design methods and widely changing real-world conditions, such as shadow and light changes, cause challenges to the extensive appointing of automatic methods. As a potential solution for such challenges, this research proposes the adaption of deep learning (DL), specifically convolutional neural networks (CNN), which has a high ability to learn features automatically, to identify damaged buildings from pre- and post-event RS imageries. Since RS data revolves around imagery, CNNs can arguably be most effective at automatically discovering relevant features, avoiding the need for feature engineering based on expert knowledge. In this work, we focus on RS imageries from orthophoto imageries for damaged-building detection, specifically for (i) background, (ii) no damage, (iii) minor damage, and (iv) debris classifications. The gist is to uncover the CNN architecture that will work best for this purpose. To this end, three CNN models, namely the twin model, fusion model, and composite model, are applied to the pre- and post-orthophoto imageries collected from the 2016 Kumamoto earthquake, Japan. The robustness of the models was evaluated using four evaluation metrics, namely overall accuracy (OA), producer accuracy (PA), user accuracy (UA), and F1 score. According to the obtained results, the twin model achieved higher accuracy (OA = 76.86%; F1 score = 0.761) compare to the fusion model (OA = 72.27%; F1 score = 0.714) and c

Predicting landslide occurrences can be difficult. However, failure to do so can be catastrophic, causing unwanted tragedies such as property damage, community displacement, and human casualties. Research into landslide susceptibility mapping (LSM) attempts to alleviate such catastrophes through the identification of landslide prone areas. Computational modelling techniques have been successful in related disaster scenarios, which motivate this work to explore such modelling for LSM. In this research, the potential of supervised machine learning and ensemble learning is investigated. Firstly, the Flexible Discriminant Analysis (FDA) supervised learning algorithm is trained for LSM and compared against other algorithms that have been widely used for the same purpose, namely Generalized Logistic Models (GLM), Boosted Regression Trees (BRT or GBM), and Random Forest (RF). Next, an ensemble model consisting of all four algorithms is implemented to examine possible performance improvements. The dataset used to train and test all the algorithms consists of a landslide inventory map of 227 landslide locations. From these sources, 13 conditioning factors are extracted to be used in the models. Experimental evaluations are made based on True Skill Statistic (TSS), the Receiver Operation characteristic (ROC) curve and kappa index. The results show that the best TSS (0.6986), ROC (0.904) and kappa (0.6915) were obtained by the ensemble model. FDA on its own seems effective at modelling landslide susceptibility from multiple data sources, with performance comparable to GLM. However, it slightly underperforms when compared to GBM (BRT) and RF. RF seems most capable compared to GBM, GLM, and FDA, when dealing with all conditioning factors.

In this research, we enhance the LEACH protocol by updating Cluster Head (CH) election probability function (thresholds). More probability was given to an out-of-service CH to be elected again. The idea is to get benefit from CH residual energy in order to extend the network lifetime. A new threshold was introduced which guarantees a non-zero probability value of a CH. We proposed a newly developed research technique to enhance the original LEACH protocol. The enhancement focuses on extending a WSN’s lifetime, and increasing its throughput. It is achieved by giving more probability to re-elect the expired CH that has been removed from CHs list because of its insufficient residual energy. Several experiments were conducted to evaluate the efficiency of our proposed MPF-LEACH approach. From the experimental results, a remarkable enhancement in the network lifetime and throughput is achieved. We have improved the election probability threshold for the original LEACH protocol by benefiting from the CH residual energy. As a result, the whole network lifetime was increased due to the extra chance that is given to a CH to be elected again.

Aerial videos captured using dynamic cameras commonly require background remodeling at every frame. In addition, camera motion and the movement of multiple objects present an unstable imaging environment with varying motion patterns. This makes detecting multiple moving objects a difficult task. In this chapter, a two-step framework, termed the motion differences of matched region-based features (MDMRBF), is presented. Firstly, each frame goes through super-pixel segmentation to produce regions where each frame is then represented as a region adjacency graph structure of visual appearance and geometric properties. This representation is important for correspondence discovery between consecutive frames based on multigraph matching. Ultimately, each region is labeled as either a background or foreground (object) using a proposed graph-coloring algorithm. Two datasets, namely (1) the DARPA-VIVID dataset and (2) self-captured videos using an unmanned aerial vehicle-mounted camera, have been used to validate the feasibility of MDMRBF. Comparison is also done with three existing detection algorithms where experiments show promising results with precision at 94%, and recall at 89%.

This study investigates the effectiveness of two sets of landslide conditioning variable(s). Fourteen landslide conditioning variables were considered in this study where they were duly divided into two sets G1 and G2. Two Support Vector Machine (SVM) classifiers were constructed based on each dataset (SVM-G1 and SVM-G2) in order to determine which set would be more suitable for landslide susceptibility prediction. In total, 160 landslide inventory datasets of the study area were used where 70% was used for SVM training and 30% for testing. The intra-relationships between parameters were explored based on variance inflation factors (VIF), Pearson's correlation and Cohen's kappa analysis. Other evaluation metrics are the area under curve (AUC).

In recent years, remote sensing researchers have investigated the use of different modalities (or combinations of modalities) for classification tasks. Such modalities can be extracted via a diverse range of sensors and images. Currently, there are no (or only a few) studies that have been done to increase the land cover classification accuracy via unmanned aerial vehicle (UAV)–digital surface model (DSM) fused datasets. Therefore, this study looks at improving the accuracy of these datasets by exploiting convolutional neural networks (CNNs). In this work, we focus on the fusion of DSM and UAV images for land use/land cover mapping via classification into seven classes: bare land, buildings, dense vegetation/trees, grassland, paved roads, shadows, and water bodies. Specifically, we investigated the effectiveness of the two datasets with the aim of inspecting whether the fused DSM yields remarkable outcomes for land cover classification. The datasets were: (i) only orthomosaic image data (Red, Green and Blue channel data), and (ii) a fusion of the orthomosaic image and DSM data, where the final classification was performed using a CNN. CNN, as a classification method, is promising due to hierarchical learning structure, regulating and weight sharing with respect to training data, generalization, optimization and parameters reduction, automatic feature extraction and robust discrimination ability with high performance. The experimental results show that a CNN trained on the fused dataset obtains better results with Kappa index of ~0.98, an average accuracy of 0.97 and final overall accuracy of 0.98. Comparing accuracies between the CNN with DSM result and the CNN without DSM result for the overall accuracy, average accuracy and Kappa index revealed an improvement of 1.2%, 1.8% and 1.5%, respectively. Accordingly, adding the heights of features such as buildings and trees improved the differentiation between vegetation specifically where plants were dense.

Assessment of the most appropriate groundwater conditioning factors (GCFs) is essential when performing analyses for groundwater potential mapping. For this reason, in this work, we look at three statistical factor analysis methods—Variance Inflation Factor (VIF), Chi-Square Factor Optimization, and Gini Importance—to measure the significance of GCFs. From a total of 15 frequently used GCFs, 11 most effective ones (i.e., altitude, slope angle, plan curvature, profile curvature, topographic wetness index, distance from river, distance from fault, river density, fault density, land use, and lithology) were finally selected. In addition, 917 spring locations were identified and used to train and test three machine learning algorithms, namely Mixture Discriminant Analysis (MDA), Linear Discriminant Analysis (LDA) and Random Forest (RF). The resultant trained models were then applied for groundwater potential prediction and mapping in the Haraz basin of Mazandaran province, Iran. MDA has been successfully applied for soil erosion and landslide mapping, but has not yet been fully explored for groundwater potential mapping (GPM). Although other discriminant methods, such as LDA, exist, MDA is worth exploring due to its capability to model multivariate nonlinear relationships between variables; it also undertakes a mixture of unobserved subclasses with regularization of non-linear decision boundaries, which could potentially provide more accurate classification. For the validation, areas under Receiver Operating Characteristics (ROC) curves (AUC) were calculated for the three algorithms. RF performed better with AUC value of 84.4%, while MDA and LDA yielded 75.2% and 74.9%, respectively. Although MDA performance is lower than RF, the result is satisfactory, because it is within the acceptable standard of environmental modeling. The outcome of factor analysis and groundwater maps emphasizes on optimization of multicolinearity factors for faster spatial modeling and provides

Ontologies play a crucial role in multiagent systems (MASs) development, especially for domain knowledge modeling, interaction specifications, and behavioral aspect representation. Domain‐specific ontologies can be developed in an ad hoc or systematic manner through the incorporation of ontology development steps on the basis of agent‐oriented methodologies. Developing such ontologies, however, is challenging because of the extensive amounts of knowledge and experience required. Moreover, since many ontologies cater for very specific domains, the question arises of whether some can be reused for faster systems development. This paper attempts to answer this question by proposing an ontology pattern classification scheme to allow the reuse of existing ontology knowledge for MAS development. Specifically, ontology patterns relevant to the design problem at hand are identified through the pattern classification scheme. These patterns are then reused and shared among agent software communities during the system development phase. The effectiveness of the proposed approach is validated using a restaurant‐finder MAS case study. Our findings suggest that utilization of the classified ontology patterns reduces development time and complexity when dealing with domain‐specific applications. The scheme also seems useful for software practitioners, where searching and reusing the patterns can easily be done during the analysis, design, and implementation of MAS development.

Food has become one of the most photographed objects since the inceptions of smart phones and social media services. Recently, the analysis of food images using object recognition techniques have been investigated to recognize food categories. It is a part of a framework to accomplish the tasks of estimating food nutrition and calories for health-care purposes. The initial stage of food recognition pipeline is to extract the features in order to capture the food characteristics. A local feature by using SURF is among the efficient image detector and descriptor. It is using fast hessian detector to locate interest points and haar wavelet for descriptions. Despite the fast computation of SURF extraction, the detector seems ineffective as it obviously detects quite a small volume of interest points on the food objects with monotonous appearance. It occurs due to 1) food has texture-less surface 2) image has small pixel dimensions, and 3) image has low contrast and brightness. As a result, the characteristics of these images that were captured are clueless and lead to low classification performance. This problem has been manifested through low production of interest points. In this paper, we propose a technique to detect denser interest points on monotonous food by increasing the density of blobs in fast hessian detector in SURF. We measured the effect of this technique by performing a comparison on SURF interest points detection by using different density of blobs detection. SURF is encoded by using Bag of Features (BoF) model and Support Vector Machine (SVM) with linear kernel adopted for classification. The findings has shown the density of interest point detection has prominent effect on the interest points detection and classification performance on the respective food categories with 86% classification accuracy on UEC100-Food dataset.

Sentiment analysis is the computational study of opinions given by the users of online media platforms e.g. Twitter, Facebook, Instagram. The output will be in the form of polarity: positive, negative or indifferent. The field has become very useful for the industry as it can feed them the information of what is sought after by their customers in a given time. It has also rapidly became a topic of interest in the research world, for its importance and subjectivity. One of the most challenging issue in sentiment analysis is sarcasm. The existence of sarcasm is mostly ignored by the researchers in the field of sentiment analysis as it is considered to be too complex. Sarcasm is what most researchers regarded as a subset of irony. It is the utterance of positive statement with negative intent. Intent is hard to detect not only for computers but also for humans. The listener is deemed to have a certain degree of background knowledge or context of what the speaker is saying to understand sarcasm. The researches that takes sarcasm into account or solely focuses on sarcasm is in the trend of using context outside the target word for sarcasm detection, and the most popular approach is deep learning. However, both deep learning and context need a lot of features. In this paper, we will look at some researches that focuses on sarcasm detection and their agreement that more than text is needed to properly detect sarcasm. Also in this paper is the trends undergone by sarcasm detection researchers and their proposed techniques.

Food object recognition has gained popularity in recent years. This can perhaps be attributed to its potential applications in fields such as nutrition and fitness. Recognizing food images however is a challenging task since various foods come in many shapes and sizes. Besides having unexpected deformities and texture, food images are also captured in differing lighting conditions and camera viewpoints. From a computer vision perspective, using global image features to train a supervised classifier might be unsuitable due to the complex nature of the food images. Local features on the other hand seem the better alternative since they are able to capture minute intricacies such as interest points and other intricate information. In this paper, two local features namely SURF (Speeded- Up Robust Feature) and MSER (Maximally Stable Extremal Regions) are investigated for food object recognition. Both features are computationally inexpensive and have shown to be effective local descriptors for complex images. Specifically, each feature is firstly evaluated separately. This is followed by feature fusion to observe whether a combined representation could better represent food images. Experimental evaluations using a Support Vector Machine classifier shows that feature fusion generates better recognition accuracy at 86.6%.

The social media services such as Facebook, Instagram and Twitter has attracted millions of food photos to be uploaded every day since its inception. Automatic analysis on food images are beneficial from health, cultural and marketing aspects. Hence, recognizing food objects using image processing and machine learning techniques has become emerging research topic. However, to represent the key features of foods has become a hassle from the immaturity of current feature representation techniques in handling the complex appearances, high deformation and large variation of foods. To employ many kinds of feature types are also infeasible as it inquire much pre-processing and computational resources for segmentation, feature representation and classification. Motivated from these drawbacks, we proposed the integration on two kinds of local feature namely Speeded-Up Robust Feature (SURF) and Scale Invariant Feature Transform (SIFT) to represent the features large variation food objects. Local invariant features have shown to be successful in describing object appearances for image classification tasks. Such features are robust towards occlusion and clutter and are also invariant against scale and orientation changes. This makes them suitable for classification tasks with little inter-class similarity and large intra-class difference. The Bag of Features (BOF) approach is employed to enhance the discriminative ability of the local features. Experimental results demonstrate impressive overall recognition at 82.38% classification accuracy from the local feature integration based on the challenging UEC-Food100 dataset. Then, we provide depth analysis on SURF and SIFT implementation to highlight the problems towards recognizing foods that need to be rectified in the future research.

Mobile application development is receiving much attention nowadays. With the enhancement of mobile application tools like an Android studio, etc. and kinds of online support, the development of the mobile application is getting easier. Indeed, mobile application development is not a trivial task. When given a particular problem, a novice mobile programmer will commonly sketch the mobile interface followed by coding. The rapid prototyping technique and trial from errors have led to issues such as poor domain understanding. We argue that a complete understanding of the domain is needed for mobile application development. Hence, requirements engineering is an important phase. This paper introduces a technique to assist mobile application development through Agent-Oriented Requirements Engineering (AORE). AORE consists of goal modelling to analyse and understand a mobile-based project. With goal modelling, AORE allows a modeller to identify and analyse the functionalities and non-functionalities of the system and present a holistic view of the proposed system. It showcases the services, operations and constraints of the proposed system. AORE is a useful part of the development phase and can complement current steps in mobile application development lifecycle.

One pressing issue in resource constrained schools is the lack of access to computer facilities for learning. A single display groupware (SDG) is explored in this paper to provide an alternative solution for this limitation where it allows multiple users to work concurrently on a single computer display. This study aims to investigate the feasibility of SDG as an educational approach in resource constrained schools. The prototype SDG system RimbaIlmu, is designed and developed. Quantitative evaluations are carried out where the results indicate its feasibility in students’ engagement and learning performance, while overcoming technical constraints. Limitation of the SDG is also presented with the intent of guiding future improvements.

This paper presents the post‐mortem report upon completion of the Long Lamai e‐commerce development project. Some weaknesses with regards to the current software modelling approach are identified and an alternative role‐based approach is proposed. We argue that the existing software modelling technique is not suitable for modelling, making it difficult to establish a good contract between stakeholders causing delays in the project delivery. The role‐based approach is able to explicitly highlight the responsibilities among stakeholders, while also forming the contract agreement among them leading towards sustainable ICT4D.

Semantic image segmentation is a fundamental yet challenging problem, which can be viewed as an extension of the conventional object detection with close relation to image segmentation and classification. It aims to partition images into non-overlapping regions that are assigned predefined semantic labels. Most of the existing approaches utilize and integrate low-level local features and high-level contextual cues, which are fed into an inference framework such as, the conditional random field (CRF). However, the lack of meaning in the primitives (i.e., pixels or superpixels) and the cues provides low discriminatory capabilities, since they are rarely object-consistent. Moreover, blind combinations of heterogeneous features and contextual cues exploitation through limited neighborhood relations in the CRFs tend to degrade the labeling performance. This paper proposes an ontology-based semantic image segmentation (OBSIS) approach that jointly models image segmentation and object detection. In particular, a Dirichlet process mixture model transforms the low-level visual space into an intermediate semantic space, which drastically reduces the feature dimensionality. These features are then individually weighed and independently learned within the context, using multiple CRFs. The segmentation of images into object parts is hence reduced to a classification task, where object inference is passed to an ontology model. This model resembles the way by which humans understand the images through the combination of different cues, context models, and rule-based learning of the ontologies. Experimental evaluations using the MSRC-21 and PASCAL VOC'2010 data sets show promising results.

Over the past two decades, the use of computer vision methods for enabling machines to recognize human action from a sequence of images, has grown as information technologies advance, and hardware availability such as cameras (especially closed circuit television) has increased. From the latter part of the 1980s till recently, computer vision has been employed for human action recognition research. Due to the volume of existing academic studies, it would be impractical to review all researches. This paper presents a brief analysis regarding the body of knowledge in Human Activity Recognition from 1987 to 2015. Bibliometric techniques based on the Science Citation Index (SCI) databases of the Web of Science are employed where 1,172 articles are critically analysed on the various aspects of publication characteristics such as authorship, countries, institutions, number of citations, and keywords. The pace of publishing in this field has shown to increase rapidly over last 20 years. By identifying the global trends in HAR research, this study is beneficial for researchers, for example, in the selection of future research topics. Similarly, policy makers can also benefit from the findings for a better understanding of how HAR develops over time.

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.

Agent technology has been used in building various domain specific applications. The agent methodologies are proposed to aid the agent developer with the introduction of techniques, terminology, notation and guidelines during the development of the agent system. Alternatively, agent patterns have been introduced by sharing the experience of engineering the agent system and allow the novices to solve the problem in a more systematic and structured way. To ease the accessibility of agent patterns, various catalogs or pattern classifications have been introduced. The pattern classification allows the user to find the patterns by organizing the patterns within a particular catalog. To date, various styles of pattern classification have been introduced. We argue that those styles are still insufficient to classify the current pools of agent patterns. This paper presents a classification scheme for agent patterns. It is an improvement of the existing pattern classifications. The improved classification was able to classify 204 agent patterns, which indirectly will ease of pattern selection in multi agent system development. In addition, we show the usage of pattern classification in determine the quality of the agent patterns. In fact, it is the first report that shows the quality of the agent patterns to date.

Image segmentation is still considered a very challenging subject despite years of research effort poured into the field. The problem is exacerbated when there is need for specific object detection. Since objects can be visually non-homogeneous, techniques that attempt to segment images into visually uniform regions using only the bottom-up cues, tend to fail. We propose a novel two-step model that incorporates both bottom-up information and top-down object constraints. Firstly, a set of uniform regions are generated using an extension of contour detection, seeded region growing, and graph-based methods. The second step applies co-occurrence constraints on the image regions in order to perceptually group regions into objects. This unsupervised segmentation process models each object using higher-level knowledge in the form of visual co-occurrences of its constituent parts. Experiments on the horse and ImageCLEF databases show that the proposed technique performs comparably well with existing state-of-the-art techniques.

Human action recognition from video has several potential to apply in different real-life applications, but the most cases in this field suffer from the variation in viewpoint. Most of published methods in this area are considered the performance of each single camera, therefore the change in the viewpoints significantly decrease the recognition rate. In this paper, multiple views are considered together and a method has proposed to recognize human action depicted in multi-view image sequences. In the first step, the border of the human body's silhouette is extracted and distance signal is calculated. In the next step, the wavelet transform is applied to extract coefficients of single-view features, and then the extracted features are combined to compose multi-view features. Finally a hierarchical classifier using support vector machine and Naïve Bayes classifiers is implemented to classify the actions. The average of overall action recognition accuracy for 12 actions using 5 different angles of views on the IXMAS dataset is 88.22. The results of experiments on the popular multi-view dataset have shown the proposed method achieves high and state-of-the-art success rates. In other word, combination of single-view extracted features from the wavelet approximation coefficients and composing the multi-view features can be used as the multi-view features. Further, the hierarchical classifier can be applied to recognize actions in multi-view human action recognition area.

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.

This paper proposes an Eye Gaze Tracking (EGT) technique using a single eye image that can be easily calibrated and mapped for Human Computer Interaction (HCI). The technique employs both geometric and trigonometric relationships to find a user's Point of Regard, followed by calculating user-dependent variables for final mapping onto a user interface (UI). Experimental results show acceptable accuracy with minimal focal errors.