An Automatic Video Timestamp Technique for Monitoring Moving Objects Using YOLOv5

Navigating long videos to find important moments can often be a complicated task. This is due to the lack of an effective video content indexing mechanism. Timestamps, time markers embedded in video streams, offer a potential solution by allowing viewers to jump directly to specific sections of interest. This study explores the role of timestamps in improving the effectiveness of monitoring objects in videos. A video timestamp-based object detection method using YOLOv5 algorithm is implemented and evaluated here. The test scenario results show that YOLOv5 produces good object detection with accurate timestamps, with an F1 score of 0.772. This implies that YOLOv5 method is effective as a timestamp method for monitoring moving objects in videos.