English static mirror for SEO/GEO · AI-assisted translation · Read Chinese original

GeoNeXt: Video Generative Models as Geometry Learner for Depth and Normal Estimation

Forum topic · 小凯 · 2026-09-01

Summary

GeoNeXt is a new framework that repurposes pretrained video generative models as a unified, data-efficient approach to monocular geometry estimation. Instead of adapting image diffusion models—which either train separate task-specific models for depth and surface normals or require large labeled datasets for joint fine-tuning—GeoNeXt formulates geometry estimation as a next-frames prediction task. This allows the model to inherit the naturally structured knowledge and richer priors of video models while jointly modeling image and geometric targets. Experiments on diverse benchmarks demonstrate state-of-the-art zero-shot monocular depth and surface normal estimation, outperforming prior task-specific and unified generative competitors with substantially less training data. Notably, GeoNeXt rivals discriminative methods trained with over 100x more data, even surpassing them on several benchmarks. Paper: arXiv 2608.28549.

Overview

  • Field: Computer Vision
  • Authors: Haosen Yang, Jifei Song, Zhensong Zhang, Xiatian Zhu, Jiankang Deng
  • Published: 2026-08-28
  • arXiv: 2608.28549

Abstract

Recent generative approaches to geometry estimation adapt pretrained image diffusion models and treat the task as image-conditioned generation. Leveraging off-the-shelf image diffusion models, they either (i) train task-specific geometry models (for depth and surface normal estimation) independently, losing the opportunity of exploring the intrinsic correlation of these geometric targets, or (ii) jointly fine-tune modified image diffusion backbones (e.g., altered self-attention), which typically demands substantial labeled data.

To overcome these limitations in a principled fashion, the authors repurpose pretrained video generative models as a unified and data-efficient framework for geometry estimation, formulated as a next-frames prediction task. The proposed method, GeoNeXt, inherits the naturally structured knowledge and richer priors of video models, and further adapts them to jointly model image and geometric targets (image ↔ geometry), enabling more efficient and effective geometry learning.

Extensive experiments validate the approach on zero-shot monocular depth and surface normal estimation across diverse datasets, outperforming prior task-specific and unified generative competitors while using substantially less training data. Notably, GeoNeXt is competitive with state-of-the-art discriminative methods trained with over 100x more data, and even outperforms them on several benchmarks.

---

*Auto-collected on 2026-09-01.*

Tags

#computer-vision#video-generation#depth-estimation#surface-normal-estimation#diffusion-models#generative-models#geometry-estimation#arxiv

This page is an English static mirror generated for search and AI citation. It may be a full translation or structured summary of the Chinese original. Canonical interactive discussion lives on the Chinese page: https://zhichai.net/topic/178634342