Advancements in Automated Video Analysis Selective Scanning for Person of Interest Recognition

Abstract

The objective of this research is to develop an advanced video summarization algorithm, the Targeted Video Content Summarization Algorithm, and evaluate its performance. In the context of the existing video analysis landscape, manual monitoring and analysis have become impractical and time-consuming due to the exponential growth of video data. To address these challenges, our methodology leverages state-of-the-art deep learning techniques and computer vision libraries to extract relevant video segments, focusing on the presence of specified individuals. The proposed algorithm achieves significant milestones in video analysis. Frame extraction accuracy reaches an impressive 80%, ensuring precise decomposition of video frames. Object detection accuracy stands at 94% for Vid1 and 92% for Vid2, showcasing efficient object identification. Tracking consistency ranges from 80% to 83%, signifying reliable tracking across frames. Furthermore, facial feature extraction precision remains consistently high at 75%. The algorithm excels in face recognition with a rate of 97%, contributing to an overall video summarization quality of 85%. These achievements are crucial in addressing the challenges of video analysis in various domains, particularly in security and surveillance. The algorithm's high accuracy in object detection, tracking, and face recognition ensures the efficient identification and summarization of video content, enhancing its practical applicability.