RGB-D Video Mirror Detection

Mingchen Xu, Peter Herbert,Yu-Kun Lai,Ze Ji,Jing Wu

IEEE Winter Conference on Applications of Computer Vision. IEEE Winter Conference on Applications of...（2025）

Cardiff Univ

被引用19|浏览24

摘要

Mirror detection aims to identify mirror areas in a scene, with recent methods either integrating depth information (RGB-D) or making use of temporal information (video). However, utilizing both data is still under-explored due to the lack of a high-quality dataset and an effective method for the RGB-D Video Mirror Detection (DVMD) problem. To the best of our knowledge, this is the first work to address the DVMD problem. To exploit depth and temporal information in mirror segmentation, we first construct a large-scale RGB-D Video Mirror Detection Dataset (DVMD-D), which contains 17977 RGB-D images from 273 diverse videos. We further develop a novel model, named DVMDNet, which can first locate the mirrors based on triple consistencies: local consistency, cross-modality consistency and global consistency, and then refine the mirror boundaries through content discontinuity, taking the temporal information within videos into account. We conduct a comparative study on the DVMD dataset, evaluating 12 state-of-the-art models (including single-image mirror detection, single-image glass detection, RGB-D mirror detection, video shadow detection, video glass detection, and video mirror detection methods). Code is available from https://github.com/UpChen/2025_DVMDNet.

更多

查看译文

关键词

Temporal Information,Depth Information,RGB-D Images,Global Consistency,Convolutional Layers,Sequencing Depth,Local Features,Global Features,Intersection Over Union,High-level Features,Single Frame,Consistent Feature,Depth Perception,Previous Frame,Depth Features,Temporal Consistency,Dynamic Scenes,Static Scenes,Long-term Characteristics,Atrous Spatial Pyramid Pooling,RGB Features,Backbone Feature,Appearance Information,Extra Input,Boundary Information,3D Reconstruction,Binary Cross Entropy,Feature Extraction Backbone