our-model-predicts-metric-scale-pixel-aligned-and-dense-dept

NaN Xue (@NanXue7)

2026-01-28 | ❤️ 139 | 🔁 10 | 💬 3

---

Our model predicts metric-scale, pixel-aligned, and dense depth maps with no missing values. Below, we present an example of estimating the height of a cylinder using the predicted depth 🔻 https://x.com/NanXue7/status/2016400933320851822/photo/1

---

🔗 링크

• https://x.com/NanXue7/status/2016400933320851822/photo/1

---

미디어

---

🔗 Related

• bringing-foundation-models-to-depth-sensing-defm-is-trained- — 도메인: Vision/3D
• lingbot-depth-masked-depth-modeling-for-spatial-perception-941785 — 도메인: Vision/3D
• what-if-your-robot-or-car-could-see-depth-more-clearly-than- — 도메인: Vision/3D
• 1n-rotary-position-embeddings-rope-are-ubiquitous-across-tra — 도메인: Vision/3D
• actionmesh — 도메인: Vision/3D

인용 트윗

Benjamin De Kraker (@BenjaminDEKR)

Interesting project. Curious if it supports metric depth (actual depth distance estimation in real numbers)

원본 트윗

Tags

AI-ML 3D-Vision