Saved in:
Bibliographic Details
Main Authors: Lyu, Wenzhou, Lin, Jialing, Ren, Wenqi, Xia, Ruihao, Qian, Feng, Tang, Yang
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2506.21034
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916812027854848
author Lyu, Wenzhou
Lin, Jialing
Ren, Wenqi
Xia, Ruihao
Qian, Feng
Tang, Yang
author_facet Lyu, Wenzhou
Lin, Jialing
Ren, Wenqi
Xia, Ruihao
Qian, Feng
Tang, Yang
contents Commercial RGB-D cameras often produce noisy, incomplete depth maps for non-Lambertian objects. Traditional depth completion methods struggle to generalize due to the limited diversity and scale of training data. Recent advances exploit visual priors from pre-trained text-to-image diffusion models to enhance generalization in dense prediction tasks. However, we find that biases arising from training-inference mismatches in the vanilla diffusion framework significantly impair depth completion performance. Additionally, the lack of distinct visual features in non-Lambertian regions further hinders precise prediction. To address these issues, we propose \textbf{DidSee}, a diffusion-based framework for depth completion on non-Lambertian objects. First, we integrate a rescaled noise scheduler enforcing a zero terminal signal-to-noise ratio to eliminate signal leakage bias. Second, we devise a noise-agnostic single-step training formulation to alleviate error accumulation caused by exposure bias and optimize the model with a task-specific loss. Finally, we incorporate a semantic enhancer that enables joint depth completion and semantic segmentation, distinguishing objects from backgrounds and yielding precise, fine-grained depth maps. DidSee achieves state-of-the-art performance on multiple benchmarks, demonstrates robust real-world generalization, and effectively improves downstream tasks such as category-level pose estimation and robotic grasping.
format Preprint
id arxiv_https___arxiv_org_abs_2506_21034
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle DidSee: Diffusion-Based Depth Completion for Material-Agnostic Robotic Perception and Manipulation
Lyu, Wenzhou
Lin, Jialing
Ren, Wenqi
Xia, Ruihao
Qian, Feng
Tang, Yang
Computer Vision and Pattern Recognition
Commercial RGB-D cameras often produce noisy, incomplete depth maps for non-Lambertian objects. Traditional depth completion methods struggle to generalize due to the limited diversity and scale of training data. Recent advances exploit visual priors from pre-trained text-to-image diffusion models to enhance generalization in dense prediction tasks. However, we find that biases arising from training-inference mismatches in the vanilla diffusion framework significantly impair depth completion performance. Additionally, the lack of distinct visual features in non-Lambertian regions further hinders precise prediction. To address these issues, we propose \textbf{DidSee}, a diffusion-based framework for depth completion on non-Lambertian objects. First, we integrate a rescaled noise scheduler enforcing a zero terminal signal-to-noise ratio to eliminate signal leakage bias. Second, we devise a noise-agnostic single-step training formulation to alleviate error accumulation caused by exposure bias and optimize the model with a task-specific loss. Finally, we incorporate a semantic enhancer that enables joint depth completion and semantic segmentation, distinguishing objects from backgrounds and yielding precise, fine-grained depth maps. DidSee achieves state-of-the-art performance on multiple benchmarks, demonstrates robust real-world generalization, and effectively improves downstream tasks such as category-level pose estimation and robotic grasping.
title DidSee: Diffusion-Based Depth Completion for Material-Agnostic Robotic Perception and Manipulation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2506.21034