This repository contains detection results exported to a Parquet file.
Each row in video_detections.parquet corresponds to a single object detection in the video.
Columns
video_id (string)
Identifier for the processed video.
frame_index (int)
Index of the sampled frame within the video.
timestamp_sec (float)
Timestamp of the frame in seconds.
class_label (string)
Predicted exterior component label.
x_min (float)
Left coordinate of bounding box.
y_min (float)
Top coordinate of bounding box.… See the full description on the dataset page:
https://huggingface.co/datasets/AdityaBaradi/CS370_HW2.
