#223 · Primary category： Computer Vision
BundleTrack
[IROS 2021] BundleTrack: 6D Pose Tracking for Novel Objects without Instance or Category-Level 3D Models
Project last updated：04/13/26
GitHub Stars
686
Forks
72
Contributors
1
License
Apache-2.0
Why we included this project
Most 6D pose trackers assume you have a CAD model of the object, at least at the category level, before you can do anything. BundleTrack drops that requirement: it tracks the full 6D pose of novel objects using only segmentation and feature extraction from deep learning, plus a memory-augmented pose graph optimizer that keeps the estimate consistent over time. The result is long-term, low-drift tracking that holds up under heavy occlusion and fast motion, and the CUDA implementation runs the whole pipeline at about 10Hz, which is fast enough for real-time robotics work. The IROS 2021 paper and code are a solid reference if you are building manipulation or AR systems and do not want to pre-train on the specific object you will be tracking.
Articles for this project
No articles for this project yet.
To suggest a topic or contribute an article, contact us.
Related projects in this category
opencv
Open Source Computer Vision Library
RuView
π RuView turns commodity WiFi signals into real-time spatial intelligence, vital sign monitoring, and presence detection — all without a single pixel of video.
PaddleOCR
Turn any PDF or image document into structured data for your AI. A powerful, lightweight OCR toolkit that bridges the gap between images/PDFs and LLMs. Supports 100+ languages.
MinerU
Transforms complex documents like PDFs and Office docs into LLM-ready markdown/JSON for your Agentic workflows.
tesseract
Tesseract Open Source OCR Engine (main repository)