AI Trainer
VideoBounding Box
The project scope involved developing a real-time computer vision system to detect fire, smoke, and human presence from image and video data. Data labeling tasks included annotating images with bounding boxes around fire, smoke, and people using LabelImg. The dataset comprised several thousand annotated images and video frames for training and testing. Model development was carried out using YOLOv9 and OpenCV for real-time processing. Quality measures included ensuring annotation consistency, validation checks, and evaluation using precision, recall, and accuracy.
The project scope involved developing a real-time computer vision system to detect fire, smoke, and human presence from image and video data. Data labeling tasks included annotating images with bounding boxes around fire, smoke, and people using LabelImg. The dataset comprised several thousand annotated images and video frames for training and testing. Model development was carried out using YOLOv9 and OpenCV for real-time processing. Quality measures included ensuring annotation consistency, validation checks, and evaluation using precision, recall, and accuracy.
2025 - 2026