7 days ago · Depth Anything is a monocular depth estimation model developed by researchers at HKU, TikTok, CUHK and ZJU. This model creates a depth representation of an ...

2 days ago · We utilize the official Marigold evaluation pipeline to evaluate the affine-invariant depth estimation checkpoints, and we use the official DSINE evaluation ...

6 days ago · Recent work showed that large diffusion models can be reused as highly precise monocular depth estimators by casting depth estimation as an image-conditional ...

2 days ago · Paper discussion, DROID-SLAM: Deep Visual SLAM for Monocular, Stereo, and RGB-D Cameras ... Single image depth estimation, Deep Multi-view stereo. Thu 10/31/2024 ...

2 days ago · Please watch the demo video in the supplementary material. Report ... Koltun, “Towards robust monocular depth estimation: Mixing datasets for zero ...

7 days ago · Our latest one-step diffusion estimation model for specular reflection removal in arbitrary-length video sequences Demo: https://huggingface.co/spaces/Stable-X/ ...

3 days ago · Unveiling BetterDepth — a plug-and-play diffusion-based refiner for zero-shot monocular depth estimation, compatible with many established depth prediction ...

3 days ago · This led us to use several monocular depth cues, such as shape from shading and shape from texture. We even tried to use gradient color change with respect ...

4 days ago · Perform a markerless kinematic analysis from multiple calibrated views as a unified workflow from an OpenPose input to an OpenSim result.

6 days ago · Reconstructing 3D indoor scenes from 2D images has always been an important task in computer vision and graphics applications. For indoor scenes ...