Violence against women and children is a public health issue of pandemic proportions. It is estimated that one in every three women worldwide has experienced physical, emotional, or sexual violence. Similarly, each year one out of two children are victims of some form of violence including domestic aggression and bullying. Due to the widespread use of the Internet and social media, women and children are now vulnerable to other types of violence such as cyber-bullying and online sexual or emotional harassment. To help alleviate this social problem, the use of computer science and related technologies has been leveraged in recent years. The Internet of Things, artificial intelligence, ubiquitous and mobile computing, pattern recognition, cloud computing and similar technologies, have been used to formulate solutions to detect and prevent violent acts against women and children. In this paper, a systematic review of some of the efforts that can help address the problem of violence against women and children is presented. This paper describes the current state-of-the-art of these contributions and identifies trends, architectures, technologies, and current open challenges. The survey was developed using a literature review of academic documents published from 2010 to 2020. The contributions were categorized into four application domains: online detection, offline detection, safety, and education. These contributions were further categorized based on the computer science approaches and technologies used: artificial intelligence, Internet of Things, and digital serious games.
8 February 2023 in Structural Control and Health Monitoring
Electrical resistance tomography (ERT) is a nondestructive evaluation technique that uses the internal conductivity variations of materials to assess structural integrity. Due to the low instrumentation required, the widespread use of ERT in the aerospace industry for monitoring the accumulation of damage in aircraft components can lead to significant reductions in inspections and maintenance costs. However, implementing the ERT method for mapping the damage state of structural components made of carbon fiber reinforced polymeric (CFRP) composites is challenging due to the inability of this method to distinguish between damage modes such as delamination and matrix cracking. This article explores the combined use of ERT and machine learning algorithms such as neural networks, random forests, k-nearest neighbors, and support vector machines to classify and characterize delamination and matrix cracking damage in CFRP laminates. Results show that the proposed classification algorithms can successfully estimate the damage severity of delaminated composites in the presence of matrix cracking. Similarly, the classification algorithms can characterize these independent damage modes with an accuracy of 95%. The algorithms showed robustness to predict the electrical resistance variations of damaged composites and characterize delamination and matrix cracking damage even when intrinsic noise was considered. Although neural networks characterized damage with the highest accuracy, these algorithms were also the most sensitive to noise. For applications where instrumentation noise cannot be completely removed from the ERT signals, the use of nearest neighbors is thus recommended.
The use of wireless sensor networks (WSN) in tracking applications is growing at a fast pace. In these applications, the sensor nodes discover, monitor and track an event or target object. A significant number of proposals relating the use of WSNs for target tracking have been published to date. However, they either focus on the tracking algorithm or on the communication protocol, and none of them address the problem integrally. In this paper, a comprehensive proposal for target detection and tracking is discussed. We introduce a tracking algorithm to detect and estimate a target location. Moreover, we introduce a low-overhead routing protocol to be used along with our tracking algorithm. The proposed algorithm has low computational complexity and has been designed considering the use of a mobile sink while generating minimal delay and packet loss. We also discuss the results of the evaluation of the proposed algorithms.
Real-time scheduling theory has shown an impressive evolution in the past few years. As a consequence of the intensive research done in this area a lot of new scheduling policies had been proposed to date. Nevertheless, just a few of such scheduling policies are available in existing real-time operating systems (RTOS). In this paper, we describe Realtss, an open source realtime scheduling simulator which is suited to simulate real-time scheduling algorithms without the need of implement them in a RTOS. Realtss is an invaluable teaching and researching tool since existing and new real-time scheduling policies can be easily evaluated.
In recent years, the rates of firearm-related violent acts worldwide have risen significantly. It is a severe social problem that compromises the safety of every individual to some extent. This situation has motivated researchers to find new ways to improve the current state-of-the-art solutions, such as automatic surveillance systems, to detect and classify the presence of firearms within a specific scene. These systems reduce the drawbacks of using direct human supervision. Among the available solutions for the classification task, the performance of Deep Learning models stands out, especially those based on Convolutional Neural Networks. Since they start learning directly from raw data (e.g., images), their learning process can be improved even further by using Transfer Learning techniques. However, the classification accuracy depends significantly on choosing the optimum set of values for the different hyperparameters composing them. Thus, this paper analyses the improvement in the performance of an image-based handgun classification algorithm when tuning its hypermeters values instead of using its default values. In this work, we evaluated the performance variation using two benchmarks Convolutional Neural Networks architectures: AlexNet and Inception V3. We obtained a maximum accuracy of 94.11% when using the Inception V3 network and transfer learning. We employed Nadam as the optimizer and a learning rate equal to 0.0001, a batch size equal 256, and a total of 13 epochs. Experimental results suggest an essential relationship between the performance of the classification model and the data set, the specific combinations of values for the selected optimizer, the batch size, and the learning rate. The obtained improvement in the accuracy was up to 10.33% after the tuning process.