PXDepth: Pixel-Space Modeling for Structure Preserving Monocular Depth Estimation

2026-08-17

AI Summary

本文针对单目深度估计中细粒度结构和物体边界保留问题，提出PXDepth模型，通过分离全局上下文建模与像素级深度预测来解决。

Abstract

Recent monocular depth estimators achieve strong zero-shot generalization, yet often struggle to preserve fine-grained structures and object boundaries. We attribute this limitation to the prevalent combination of large-patch ViT encoders and convolutional decoders, as coarse tokenization can weaken pixel-level cues that upsampling cannot fully recover. To address this issue, we propose PXDepth, a discriminative monocular depth model that separates global context modeling from pixel-level depth prediction. Specifically, a large-patch ViT captures global scene context, while a pixel-space predictor composed of Context-Modulated Pixel Transformer blocks maintains high-resolution spatial representations throughout depth estimation. This design preserves fine structures and sharp boundaries without sacrificing global depth consistency. Across diverse zero-shot benchmarks, PXDepth combines faithful local geometry with competitive global depth accuracy while remaining efficient at inference. Our code and model are available at https://yuanzhy29.github.io/PXDepth-Page/.

Problem

• monocular depth estimation
• fine-grained structures
• object boundaries

Innovation

• Pixel-Space Modeling
• Context-Modulated Pixel Transformer
• Monocular Depth Estimation
• Fine-Grained Structures