Tracklet Parser

A parser for tracklet labels in KITTI Raw Format 1.0 created by the Computer Vision Annotation Tool (CVAT).

Installation

Assuming that you already have a working Python environment, you can install all necessary packages with

python -m pip install .

Usage

Creating a folder with text files in KITTI format. Each text file contains the labeling information of its corresponding recording.

from tracklet_parser.tracklet import Tracklet
from tracklet_parser.tracklet_parser import TrackletParser

def main():
    tracklet_labels: str = "C:\\Foo\\tracklet_labels.xml"
    frame_list: str = "C:\\Foo\\frame_list.txt"
    output_dir: str = "C:\\Foo\\Bar"

    tracklets: List[Tracklet] = TrackletParser.parse_tracklet_xml(tracklet_labels)
    # Create n label text files in C:\Foo\Bar for n (labeled) recordings
    TrackletParser.convert_tracklets_to_kitti(tracklets, frame_list, output_dir)

if __name__ == "__main__":
    main()