Locating a missing child or elderly person in a large gathering through face recognition in videos is still challenging because of various dynamic factors. In this paper, we present an intelligent mechanism for tracking missing persons in an unconstrained large gathering scenario of Al-Nabawi Mosque, Madinah, KSA. The proposed mechanism in this paper is unique in two aspects. First, there are various proposals existing in the literature that deal with face detection and recognition in high-quality images of a large crowd but none of them tested tracking of a missing person in low resolution images of a large gathering scenario. Secondly, our proposed mechanism is unique in the sense that it employs four phases: (a) report missing person online through web and mobile app based on spatio-temporal features; (b) geo fence set estimation for reducing search space; (c) face detection using the fusion of Viola Jones cascades LBP, CART, and HAAR to optimize the results of the localization of face regions; and (d) face recognition to find a missing person based on the profile image of reported missing person. The overall results of our proposed intelligent tracking mechanism suggest good performance when tested on a challenging dataset of 2208 low resolution images of large crowd gathering.
The probability of losing vulnerable companions, such as children or older ones, in large gatherings is high, and their tracking is challenging. We proposed a novel integration of face-recognition algorithms with a soft voting scheme, which was applied, on low-resolution cropped images of detected faces, in order to locate missing persons in a challenging large-crowd gathering. We considered the large-crowd gathering scenarios at Al Nabvi mosque Madinah. It is a highly uncontrolled environment with a low-resolution-images data set gathered from moving cameras. The proposed model first performs real-time face-detection from camera-captured images, and then it uses the missing person’s profile face image and applies well-known face-recognition algorithms for personal identification, and their predictions are further combined to obtain more mature prediction. The presence of a missing person is determined by a small set of consecutive frames. The novelty of this work lies in using several recognition algorithms in parallel and combining their predictions by a unique soft-voting scheme, which in return not only provides a mature prediction with spatio-temporal values but also mitigates the false results of individual recognition algorithms. The experimental results of our model showed reasonably good accuracy of missing person’s identification in an extremely challenging large-gathering scenario.
The advent of the new millennium, with the promises of the digital age and space technology, favors humankind in every perspective. The technology provides us with electric power and has infinite use in multiple electronic accessories. The electric power produced by different sources is distributed to consumers by the transmission line and grid stations. During the electric transmission from primary sources, there are various methods by which to commit energy theft. Energy theft is a universal electric problem in many countries, with a possible loss of billions of dollars for electric companies. This energy contention is deep rooted, having so many root causes and rugged solutions of a technical nature. Advanced Metering Infrastructure (AMI) is introduced with no adequate results to control and minimize electric theft. Until now, so many techniques have been applied to overcome this grave problem of electric power theft. Many researchers nowadays use machine learning algorithms, trying to combat this problem, giving better results than previous approaches. Random Forest (RF) classifier gave overwhelmingly good results with high accuracy. In our proposed solution, we use a novel Convolution Neural Network (CNN) with RUSBoost Manta Ray Foraging Optimization (rus-MRFO) and RUSBoost Bird Swarm Algorithm (rus-BSA) models, which proves to be very innovative. The accuracy of our proposed approaches, rus-MRFO and rus-BSA, are 91.5% and a 93.5%, respectively. The proposed techniques have shown promising results and have strong potential to be applied in future.
Machine learning has the potential to predict unseen data and thus improve the productivity and processes of daily life activities. Notwithstanding its adaptiveness, several sensitive applications based on such technology cannot compromise our trust in them; thus, highly accurate machine learning models require reason. Such models are black boxes for end-users. Therefore, the concept of interpretability plays the role if assisting users in a couple of ways. Interpretable models are models that possess the quality of explaining predictions. Different strategies have been proposed for the aforementioned concept but some of these require an excessive amount of effort, lack generalization, are not agnostic and are computationally expensive. Thus, in this work, we propose a strategy that can tackle the aforementioned issues. A surrogate model assisted us in building interpretable models. Moreover, it helped us achieve results with accuracy close to that of the black box model but with less processing time. Thus, the proposed technique is computationally cheaper than traditional methods. The significance of such a novel technique is that data science developers will not have to perform strenuous hands-on activities to undertake feature engineering tasks and end-users will have the graphical-based explanation of complex models in a comprehensive way—consequently building trust in a machine.
5 January 2017 in Journal of Basic & Applied Sciences
Underwater wireless sensor network is an emerging wireless networking technology (UWSN). UWSN has various applications for example it can be used for monitoring seismic activities, underwater animal, pipeline etc. UWSN face challenges in their MAC later operations. Different energy efficient MAC protocols have been proposed for underwater wireless sensor networks (UWSN) to overcome the problem of propagation delays which is inherent in underwater acoustic networks. In this paper, we study the energy efficient MAC protocols including EE-MACU, R-MAC and T-Lohi. We classify UWSN MAC protocols into two broad categories contention free and contention based and we further categorize contention based protocol. We analyze and compare key UWSN MAC protocols based on certain parameters and suggest their suitability in various scenarios.
30 June 2017 in Sukkur IBA Journal of Computing and Mathematical Sciences
In today’s world the rapid advancements in Micro-Electromechanical Systems (MEMS) and Nano technology have improved almost all the aspects of daily life routine with the help of different smart devices such as smart phones, compact electronic devices etc. The prime example of these emerging developments is the development of wireless sensors for healthcare procedures. One kind of these sensors is wearable bio-sensors. In this paper, the technologies of two types of bio-sensors (ECG, EMG) are investigated and also compared with traditional ECG, EMG equipment. We have taken SHIMMERTM wireless sensor platform as an example of wearable biosensors technology. We have investigated the systems developed for analysis techniques with SHIMMERTM ECG and EMG wearable bio-sensors and these biosensors are used in continuous remote monitoring. For example, applications in continuous health monitoring of elderly people, critical chronic patients and Fitness & Fatigue observations. Nevertheless, early fall detection in older adults and weak patients, treatment efficacy assessment. This study not only provides the basic concepts of wearable wireless bio-sensors networks (WBSN), but also provides basic knowledge of different sensor platforms available for patient’s remote monitoring. Also various healthcare applications by using bio-sensors are discussed and in last comparison with traditional ECG and EMG is presented.