Computer Vision Based Algorithms for Detecting and Classification of Activities for Fall Recognition on Real Time Video

Falls among the elderly population present significant risks, including serious injuries and even death. Prompt detection of falls is crucial for immediate assistance and notification to relevant parties. However, current fall detection systems face challenges and privacy concerns. To address this, a new system is proposed using a Convolutional Neural Network (CNN) and optimizing Optical Flow images through preprocessing in spatial and frequency domains. The system incorporates the Viola-Jones algorithm for human face and action recognition. It is designed for detecting falls in home or room environments, covering various postures - standing, sitting, walking, and falling. Surveillance cameras in public spaces capture relevant data for training the model. The system achieves impressive accuracy and precision, highlighting its effectiveness in fall detection applications.