Depth Pro 2: Sharp and Temporally Consistent Metric Monocular Depth
We present Depth Pro 2, an approach to estimating dense metric depth for images and video at full resolution. By analyzing prior methods for monocular depth estimation, we find that they commonly trade off temporal consistency and boundary sharpness: multi-frame methods favor consistency, single-frame methods favor sharpness. To overcome this trade-off, we train a single, scalable network that processes both images and video and predicts depth maps at native resolution. Depth Pro 2 produces depth maps that are metrically accurate, sharp, and temporally consistent all at once, outperforming…
Project page: https://depth-anything-3.github.io/ Paper: https://arxiv.org/pdf/2511.10647 Demo: https://huggingface.co/spaces/depth-anything/depth-anything-3 Github: https://github.com/ByteDance-Seed/depth-anything-3 Depth Anything 3, a single transformer model trained exclusively for joint any-view depth and pose estimation via a specially chosen ray representation. Depth Anything 3 reconstructs the visual space, producing consistent depth and ray maps that can be fused into accurate…