Car-parts detection and segmentation index from RAV4 dashcam-style video frames, using a YOLO26-seg model trained on the Carparts Segmentation dataset (weights/best.pt).

	Dataset schema (carparts_detections.parquet)

Column Type Description

video_id string Source video id (e.g. "rav4_video_1")

frame_index int 1-based frame number

timestamp_sec float Time in seconds

class_label string Car-part class (trunk, wheel, bumper, glass… See the full description on the dataset page: https://huggingface.co/datasets/ArdaDinc/rav4-carparts-detections.