Focusable Monocular Depth Estimation

2026-05-12

AI Summary

Existing monocular depth estimation methods treat all pixels uniformly, failing to meet user-specified accuracy requirements for regions of interest. This work proposes a Focusable Depth Estimation (FDE) task, where bounding box or textual prompts guide the model to prioritize depth accuracy, boundary sharpness, and global geometric consistency in target regions. We introduce a novel region-aware depth estimation paradigm, featuring a prompt-conditioning mechanism and a Multi-Scale Spatial Alignment (MSSA) fusion module. Built upon SAM3 and Depth Anything architectures, our approach enables dense prompt injection that balances localized focus with global structure preservation. Evaluated on our newly curated FDE-Bench benchmark, the proposed method significantly outperforms global fine-tuning baselines, particularly in foreground regions and object boundaries. Ablation studies confirm the critical role of spatial alignment in achieving these gains.

Abstract

Monocular depth foundation models generalize well across scenes, yet they are typically optimized with uniform pixel-wise objectives that do not distinguish user-specified or task-relevant target regions from the surrounding context. We therefore introduce Focusable Monocular Depth Estimation (FDE), a region-aware depth estimation task in which, given a specified target region, the model is required to prioritize foreground depth accuracy, preserve sharp boundary transitions, and maintain coherent global scene geometry. To prioritize task-critical region modeling, we propose FocusDepth, a prompt-conditioned monocular relative depth estimation framework that guides depth modeling to focus on target regions via box/text prompts. The core Multi-Scale Spatial-Aligned Fusion (MSSA) in FocusDepth spatially aligns multi-scale features from Segment Anything Model 3 to the Depth Anything family and injects them through scale-specific, gated conditional fusion. This enables dense prompt cue injection without disrupting geometric representations, thereby endowing the depth estimation model with focused perception capability. To study FDE, we establish FDE-Bench, a target-centric monocular relative depth benchmark built from image-target-depth triplets across five datasets, containing 252.9K/72.5K train/val triplets and 972 categories spanning real-world and embodied simulation environments. On FDE-Bench, FocusDepth consistently improves over globally fine-tuned DA2/DA3 baselines under both box and text prompts, with the largest gains appearing in target boundary and foreground regions while preserving global scene geometry. Ablations show that MSSA's spatial alignment is the key design factor, as disrupting prompt-geometry correspondence increases AbsRel by up to 13.8%.

Problem

• Monocular Depth Estimation
• Region-aware
• Focusable Depth
• Target Region
• Depth Accuracy

Innovation

• Focusable Monocular Depth Estimation
• Prompt-conditioned Depth Estimation
• Multi-Scale Spatial-Aligned Fusion
• Region-aware Depth Modeling
• FDE-Bench