analiza aiqbench
PXDepth addresses the structural degradation common in monocular depth estimation by decoupling global context modeling from high-resolution pixel-level prediction using transformer blocks. This architecture successfully maintains sharp object boundaries and fine-grained geometry without compromising zero-shot generalization or inference efficiency, offering a significant improvement over standard ViT-convolution hybrid approaches.
PXDepth: Pixel-Space Modeling for Structure Preserving Monocular Depth Estimation
arXiv:2608.16984v1 Announce Type: cross 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/.
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 this https URL.
Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Graphics (cs.GR)
arXiv:2608.16984 [cs.CV]
(or arXiv:2608.16984v1 [cs.CV] for this version)
https://doi.org/10.48550/arXiv.2608.16984 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 235 cuvinte
Subiecte conexe
OpenDepth: Repurposing Image Editing Priors for Robust Open-World Monocular Depth Estimation
arXiv · 14 oct. 2026
PhysReflect: Geometry and Perception Guided Diffusion for Physically-Plausible Mirror Reflections
arXiv · 22 sept. 2026
Echoverse: Deep, evolving environments for computer-use agents
Microsoft Research · 30 iul. 2026
Game Design Creativity evals
LessWrong · 12 oct. 2026