Robust Depth Super-Resolution via Adaptive Diffusion Sampling

2026-02-10

AI Summary

This work addresses the challenge of robustly reconstructing depth maps from low-resolution inputs under severe or unknown degradation. To this end, the authors propose AdaDS, a novel framework that introduces adaptive diffusion sampling to depth super-resolution for the first time. Built upon a pre-trained diffusion model, AdaDS leverages its contractive property and incorporates uncertainty estimation to dynamically select the starting timestep of the reverse diffusion process. It further injects tailored noise to steer intermediate samples toward high-probability regions of the target posterior distribution. Notably, AdaDS requires no prior knowledge of the degradation process, enabling zero-shot generalization across diverse and unknown degradations. Extensive experiments on both real-world and synthetic datasets demonstrate that AdaDS significantly outperforms existing methods, achieving superior robustness and reconstruction quality.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Deep Generative Models & Autoencoders
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

We propose AdaDS, a generalizable framework for depth super-resolution that robustly recovers high-resolution depth maps from arbitrarily degraded low-resolution inputs. Unlike conventional approaches that directly regress depth values and often exhibit artifacts under severe or unknown degradation, AdaDS capitalizes on the contraction property of Gaussian smoothing: as noise accumulates in the forward process, distributional discrepancies between degraded inputs and their pristine high-quality counterparts diminish, ultimately converging to isotropic Gaussian prior. Leveraging this, AdaDS adaptively selects a starting timestep in the reverse diffusion trajectory based on estimated refinement uncertainty, and subsequently injects tailored noise to position the intermediate sample within the high-probability region of the target posterior distribution. This strategy ensures inherent robustness, enabling generative prior of a pre-trained diffusion model to dominate recovery even when upstream estimations are imperfect. Extensive experiments on real-world and synthetic benchmarks demonstrate AdaDS's superior zero-shot generalization and resilience to diverse degradation patterns compared to state-of-the-art methods.

Problem

• Depth Super-Resolution
• Robustness
• Degradation
• Generalization
• Diffusion Models

Innovation

• adaptive diffusion sampling
• depth super-resolution
• generative prior
• zero-shot generalization
• robust recovery