preprint-of-today-jing-et-al-stereo-any-video

Kwang Moo Yi (@kwangmoo_yi)

2025-03-10 | ❤️ 193 | 🔁 32

---

Preprint of today: Jing et al., “Stereo Any Video: Temporally Consistent Stereo Matching” — https://tomtomtommi.github.io/StereoAnyVideo/

Not everything has to be diffusion-based: Use off-the-shelf monocular video depth estimator (diffusion :() + RAFT-like architecture. Cost volumes strikes back! https://x.com/kwangmoo_yi/status/1899172898579108307/video/1

🔗 원본 링크

• https://tomtomtommi.github.io/StereoAnyVideo/
• https://x.com/kwangmoo_yi/status/1899172898579108307/video/1

미디어

---

🔗 Related

Auto-generated - needs manual review

Tags

domain-vision-3d domain-genai domain-visionos