DepthFocus: Controllable Depth Estimation for See-Through Scenes

TL;DR

DepthFocus introduces a controllable depth estimation model using a steerable transformer that actively adjusts focus to resolve layered ambiguities in transparent scenes, outperforming existing methods.

Contribution

We propose DepthFocus, a novel steerable transformer that dynamically modulates depth perception based on target focus, enabling active, intent-driven 3D perception in complex scenes.

Findings

Achieves state-of-the-art results on multiple benchmarks.

Effectively resolves depth ambiguities in transparent and reflective scenes.

Maintains high precision in opaque regions while focusing on target depths.

Abstract

Depth in the real world is rarely singular. Transmissive materials create layered ambiguities that confound conventional perception systems. Existing models remain passive; conventional approaches typically estimate static depth maps anchored to the nearest surface, and even recent multi-head extensions suffer from a representational bottleneck due to fixed feature representations. This stands in contrast to human vision, which actively shifts focus to perceive a desired depth. We introduce \textbf{DepthFocus}, a steerable Vision Transformer that redefines stereo depth estimation as condition-aware control. Instead of extracting fixed features, our model dynamically modulates its computation based on a physical reference depth, integrating dual conditional mechanisms to selectively perceive geometry aligned with the desired focus. Leveraging a newly curated large-scale synthetic dataset, \textbf{DepthFocus} achieves state-of-the-art results across all evaluated benchmarks, including both standard single-layer and complex multi-layered scenarios. While maintaining high precision in opaque regions, our approach effectively resolves depth ambiguities in transparent and reflective scenes by selectively reconstructing geometry at a target distance. This capability enables robust, intent-driven perception that significantly outperforms existing multi-layer methods, marking a substantial step toward active 3D perception. \noindent \textbf{Project page}: \href{https://junhong-3dv.github.io/depthfocus-project/}{\textbf{this https URL}}.