There are 5 columns: Video ID (VID), Frame Index, Class Label, Bounding Box, Confidence Score
Each row represents 1 frame, labeled 0-930 in order of the video.
Each Class Label contains a list of what elements were detected in that frame
The Bounding Boxes are lists of lists containing the xyxy information for each element, in index order, respectively
The Confidence Score is the confidence of each detected class from Class Label
There are 23 possible class labels of different exterior car… See the full description on the dataset page:
https://huggingface.co/datasets/mg5879/AI-HW2-CarParts-Detection-Results.
