ARDepth: Auto-regressive Monocular Depth Estimation with Progressive Visual Conditioning

2026-07-14

AI Summary

Existing monocular depth estimation methods struggle to simultaneously capture the piecewise nature of scene geometry and its scale-dependent characteristics, often resulting in an imbalance between global structure and local detail. To address this, this work proposes an autoregressive depth generation paradigm that formulates depth prediction as a resolution-progressive generative process. The approach introduces a Scale-Progressive Conditioning (SPC) mechanism and a Semantic-Aware Guidance (SAG) strategy to effectively integrate multi-scale visual features with high-level semantic priors. Evaluated across multiple benchmarks, the method achieves state-of-the-art performance, producing depth maps that exhibit strong structural consistency and faithful detail recovery across scales.

Abstract

Diffusion models have recently become the dominant paradigm for monocular depth estimation (MDE). However, they implicitly assume that depth can be recovered as a globally smooth field through iterative denoising, which does not explicitly reflect the piecewise and scale-dependent organization of scene geometry. In practice, geometric structure emerges progressively across spatial scales, where coarse layout, surfaces, and boundaries are constructed in a hierarchical manner. Motivated by this observation, we introduce ARDepth, which formulates depth estimation as structured auto-regressive generation. Instead of recovering depth through global refinement, ARDepth progressively constructs depth representations as spatial resolution increases. To support this generative process, we introduce Scale-Progressive Conditioning (SPC) to inject multi-scale visual features at each generation stage, and Semantic-Aware Guidance (SAG) to provide scene-level semantic priors that enhance global structural consistency. Together, these designs enable the model to capture fine-grained local details while maintaining coherent global geometry. Empirical results demonstrate that our approach achieves strong performance and produces structurally consistent depth predictions across scales, validating auto-regressive generation as a promising alternative paradigm for geometric modeling.

Problem

• monocular depth estimation
• scene geometry
• spatial scales
• structured generation
• geometric consistency

Innovation

• auto-regressive generation
• monocular depth estimation
• scale-progressive conditioning
• semantic-aware guidance
• structured geometric modeling