TrackCraft3R: Repurposing Video Diffusion Transformers for Dense 3D Tracking

Jisu Nam, Jahyeok Koo, Soowon Son, Jaewoo Jung, Honggyu An, Junhwa Hur, Seungryong Kim

arXiv:2605.12587 · 2026-05-14 공개 · arXiv · PDF

lora-finetuning monocular-video spatio-temporal-priors motion-priors temporal-rope pointmap-prediction video-dits dense-tracking

Abstract

Dense 3D tracking from monocular video is fundamental to dynamic scene understanding. While recent 3D foundation models provide reliable per-frame geometry, recovering object motion in this geometry remains challenging and benefits from strong motion priors learned from real-world videos. Existing 3D trackers either follow iterative paradigms trained from scratch on synthetic data or fine-tune 3D reconstruction models learned from static multi-view images, both lacking real-world motion priors. Pre-trained video diffusion transformers (video DiTs) offer rich spatio-temporal priors from internet-scale videos, making them a promising foundation for 3D tracking. However, their frame-anchored formulation, which generates each frame's content, is fundamentally mismatched with reference-anchored dense 3D tracking, which must follow the same physical points from a reference frame across time. We present TrackCraft3R, the first method to repurpose a video DiT as a feed-forward dense 3D tracker. Given a monocular video and its frame-anchored reconstruction pointmap, TrackCraft3R predicts a reference-anchored tracking pointmap that follows every pixel of the first frame across time in a single forward pass, along with its visibility. We achieve this through two designs: (i) a dual-latent representation that uses per-frame geometry latents and reference-anchored track latents as dense queries, and (ii) temporal RoPE alignment, which specifies the target timestamp of each track latent. Together, these designs convert the per-frame generative paradigm of video DiTs into a reference-anchored tracking formulation with LoRA fine-tuning. TrackCraft3R achieves state-of-the-art performance on standard sparse and dense 3D tracking benchmarks, while running 1.3x faster and using 4.6x less peak memory than the strongest prior method. We further demonstrate robustness to large motions and long videos.

한국어 요약

📋 한 줄 요약

**[3D 추적 / 비디오 디퓨전]** 비디오 디퓨전 트랜스포머의 시공간 사전을 활용해 단안 비디오에서 모든 픽셀의 dense 3D 추적을 단일 forward로 수행하는 TrackCraft3R 제안.

🎯 핵심 기여도

💡 핵심 아이디어

인터넷 규모 비디오로 학습된 비디오 디퓨전 트랜스포머는 풍부한 실세계 motion prior를 갖고 있지만 frame-anchored 생성 형식 때문에 3D 추적에 부적합하다. 표현 형식만 reference-anchored로 재배치하면 그 prior를 직접 추적에 활용할 수 있다.

🔬 기술적 접근법

📊 주요 결과

💭 의의 및 한계

**의의**: 비디오 디퓨전 사전을 생성이 아닌 추적·이해 태스크로 재활용하는 일반 청사진 제시. **한계**: 비디오 DiT의 사전 학습 가용성에 의존, 극단적 폐색·매우 빠른 비강체 운동에서의 한계는 추가 검증 필요.

🚀 실용적 활용