UG Research Intern, Computer Vision Lab, IIIT Delhi
CVATVideoObject Detection
I curated and annotated a custom dataset of over 10,000 wave-cycle frames for the Surf-Assist project using object detection labeling techniques. I designed and trained multiple YOLO-based detectors and evaluated them with tracking algorithms on real-world surf video datasets. I benchmarked and validated model performance using standard MOT metrics to ensure reproducibility and accuracy.
• Labeled video dataset frames to identify waves and surfers.
• Used CVAT and OpenCV for annotation and preprocessing.
• Ensured balanced class representation and supported 91% tracking accuracy.
• Collaborated in agile sprints for efficient annotation and model improvement.
I curated and annotated a custom dataset of over 10,000 wave-cycle frames for the Surf-Assist project using object detection labeling techniques. I designed and trained multiple YOLO-based detectors and evaluated them with tracking algorithms on real-world surf video datasets. I benchmarked and validated model performance using standard MOT metrics to ensure reproducibility and accuracy. • Labeled video dataset frames to identify waves and surfers. • Used CVAT and OpenCV for annotation and preprocessing. • Ensured balanced class representation and supported 91% tracking accuracy. • Collaborated in agile sprints for efficient annotation and model improvement.
2024 - 2025