Image Annotation & Classification Project (YOLO-format Object Detection)
I participated in a real-world autonomous driving scene data annotation project, where I used YOLO-format bounding box labels to annotate 60 images and classify objects into 6 distinct classes. The work focused on high-quality image annotation for object detection and scene understanding, specifically targeting vehicles, pedestrians, cyclists, traffic signs, traffic lights, and road obstacles. The project ensured standardized and clean annotation formats, making the dataset compatible with popular computer vision frameworks. • Produced 195 normalized bounding boxes across two high-productivity annotation sessions • Developed expertise in bounding box normalization following YOLO standards • Demonstrated attention to rare classes and class imbalance challenges • Assisted in preparing recommendations for enhancing dataset quality and diversity