paper-with-me

홈 › Papers

Seeing Through Clutter: Structured 3D Scene Reconstruction via Iterative Object Removal

2026-02-03 · Rio Aguina-Kang, Kevin James Blackburn-Matzen, Thibault Groueix, Vladimir Kim, Matheus Gadelha arxiv

We present SeeingThroughClutter, a method for reconstructing structured 3D representations from single images by segmenting and modeling objects individually. Prior approaches rely on intermediate tasks such as semantic segmentation and depth estimation, which often underperform in complex scenes, particularly in the presence of occlusion and clutter. We address this by introducing an iterative object removal and reconstruction pipeline that decomposes complex scenes into a sequence of simpler subtasks. Using VLMs as orchestrators, foreground objects are removed one at a time via detection, segmentation, object removal, and 3D fitting. We show that removing objects allows for cleaner segmentations of subsequent objects, even in highly occluded scenes. Our method requires no task-specific training and benefits directly from ongoing advances in foundation models. We demonstrate stateof-the-art robustness on 3D-Front and ADE20K datasets. Project Page: https://rioak.github.io/seeingthroughclutter/

📄 PDF Abstract BibTeX arXiv:2602.04053

Code (0)

등록된 구현이 없습니다.

Tasks

Semantic SegmentationDepth Estimation

Similar Papers 제목 키워드 기반

HyReach: Vision-Guided Hybrid Manipulator Reaching in Unseen Cluttered Environments

2026-03-22 · Shivani Kamtikar, Kendall Koe, Justin Wasserman, Samhita Marri 외 arxiv

As robotic systems increasingly operate in unstructured, cluttered, and previously unseen environments, there is a growing need for manipulators that combine compliance, adaptability, and precise control. This work prese…

Motion Planning

Simulation-Ready Cluttered Scene Estimation via Physics-aware Joint Shape and Pose Optimization

2026-02-23 · Wei-Cheng Huang, Jiaheng Han, Xiaohan Ye, Zherong Pan 외 arxiv

Estimating simulation-ready scenes from real-world observations is crucial for downstream planning and policy learning tasks. Regretfully, existing methods struggle in cluttered environments, often exhibiting prohibitive…

Seeing Isn't Orienting: A Cognitively Grounded Benchmark Reveals Systematic Orientation Failures in MLLMs Supplementary

2026-03-12 · Nazia Tasnim, Keanu Nichols, Yuting Yang, Nicholas Ikechukwu 외 arxiv

Humans learn object orientation progressively, from recognizing which way an object faces, to mentally rotating it, to reasoning about orientations between objects. Current vision-language benchmarks largely conflate ori…

Scene UnderstandingObject Recognition

Scene and Human in One World: Reconstruction in a Feedforward Pass

2026-06-26 · Boao Shi, Qiao Feng, Yiming Huang, Lingjie Liu arxiv

Reconstructing humans in dynamic scenes from moving monocular cameras remains challenging due to scale ambiguity, human-scene misalignment, and occlusion interference. Rather than treating human mesh recovery and scene r…

Human Mesh Recovery

Seeing Glass: Joint Point Cloud and Depth Completion for Transparent Objects

2021-09-30 · Haoping Xu, Yi Ru Wang, Sagi Eppel, Alàn Aspuru-Guzik 외

The basis of many object manipulation algorithms is RGB-D input. Yet, commodity RGB-D sensors can only provide distorted depth maps for a wide range of transparent objects due light refraction and absorption. To tackle t…

Depth CompletionTransparent objects