Domestic Child Physical Abuse Detection using Machine Learning

Abstract

Child physical abuse is a distressing social issue that requires timely intervention to protect vulnerable children. In this work, the main detection for child physical abuse is focused on slapping action. Slapping is still a common form of physical violence in a variety of contexts, including domestic violence, child abuse, and bullying. The ability to detect and recognise slapping incidents automatically can help prevent and address such harmful behaviours. Moreover, this work aims to develop an effective and automated system for detecting domestic child physical abuse using machine learning techniques. This work focusing on child physical abuse which on slapping detection using machine learning utilizes several libraries, including MediaPipe, OpenCV, NumPy, matplotlib, and math. Furthermore, the uses of OpenCV as a computer vision library, to pre-process the collected data. This may involve tasks such as video frame extraction, resizing, and noise reduction to enhance the quality of the input data. Furthermore, the MediaPipe library uses, which provides a collection of pre-trained machine learning models for a variety of computer vision tasks, including pose estimation and hand tracking. Moreover, uses of the libraries NumPy, matplotlib, and math to assist with data manipulation, visualization, and calculations within the application. In addition, the dataset is used to train a deep learning model, such as a convolutional neural network (CNN), to learn the visual patterns associated with slapping gestures and then a separate set of videos is used for testing to evaluate the performance of the proposed slapping detection system. Based on the experiment, the testing process has yielded a result indicating an accuracy rate of 90% which it has been classified as a slapping incident with a high level of certainty. Besides, there is a limitation in the system wherein if there are multiple individuals in the video and only one is not facing the camera, it may cause confusion and ambiguity in determining which person is performing the slapping action. The experimental results indicate that the suggested technique is effective at detecting slapping gestures