CVPR 2023 BDD100K Challenges

multi-object tracking and object detection for autonomous driving

Participant in the CVPR 2023 BDD100K Challenges (object tracking, autonomous driving).

Overview

Multi-Object Tracking to Guide Autonomous Driving through Machine Learning Algorithms

This project develops detection and tracking pipelines for autonomous driving scenarios on the BDD100K benchmark.

Key Components

  • Object detection: state-of-the-art detectors including YOLO and DINO
  • Multi-object tracking: an MOT pipeline using state-of-the-art machine learning models
  • Results: top 20% in Multi-Object Tracking (MOT), 1st place in Object Detection (OD)
Tracking visualization. To be replaced with a figure from the project.