# Grounding by Remembering: Cross-Scene and In-Scene Memory for 3D Functional Affordances

arXiv: https://arxiv.org/abs/2605.11616 · HTML: https://mlsanity.com/paper/2605.11616/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 20 of its 24 references that are in the map.
- First posted on arXiv 12 May 2026.

## Builds on (most cited first, 20 of 20)

- [Segment Anything](https://mlsanity.com/paper/2304.02643/) (2023)
- [SAM 2: Segment Anything in Images and Videos](https://mlsanity.com/paper/2408.00714/) (2024)
- [SAM 3: Segment Anything with Concepts](https://mlsanity.com/paper/2511.16719/) (2025)
- [OpenScene: 3D Scene Understanding with Open Vocabularies](https://mlsanity.com/paper/2211.15654/) (2022)
- [LERF: Language Embedded Radiance Fields](https://mlsanity.com/paper/2303.09553/) (2023)
- [ScanRefer: 3D Object Localization in RGB-D Scans using Natural Language](https://mlsanity.com/paper/1912.08830/) (2019)
- [ConceptFusion: Open-set Multimodal 3D Mapping](https://mlsanity.com/paper/2302.07241/) (2023)
- [OpenMask3D: Open-Vocabulary 3D Instance Segmentation](https://mlsanity.com/paper/2306.13631/) (2023)
- [Visual Prompting via Image Inpainting](https://mlsanity.com/paper/2209.00647/) (2022)
- [SegGPT: Segmenting Everything In Context](https://mlsanity.com/paper/2304.03284/) (2023)
- [Where2Act: From Pixels to Actions for Articulated 3D Objects](https://mlsanity.com/paper/2101.02692/) (2021)
- [3D AffordanceNet: A Benchmark for Visual Object Affordance Understanding](https://mlsanity.com/paper/2103.16397/) (2021)
- [OpenIns3D: Snap and Lookup for 3D Open-vocabulary Instance Segmentation](https://mlsanity.com/paper/2309.00616/) (2023)
- [Segment3D: Learning Fine-Grained Class-Agnostic 3D Segmentation without Manual Labels](https://mlsanity.com/paper/2312.17232/) (2023)
- [Learning 2D Invariant Affordance Knowledge for 3D Affordance Grounding](https://mlsanity.com/paper/2408.13024/) (2024)
- [3D-GRES: Generalized 3D Referring Expression Segmentation](https://mlsanity.com/paper/2407.20664/) (2024)
- [Functionality understanding and segmentation in 3D scenes](https://mlsanity.com/paper/2411.16310/) (2024)
- [AffordBot: 3D Fine-grained Embodied Reasoning via Multimodal Large Language Models](https://mlsanity.com/paper/2511.10017/) (2025)
- [MVGGT: Multimodal Visual Geometry Grounded Transformer for Multiview 3D Referring Expression Segmentation](https://mlsanity.com/paper/2601.06874/) (2026)
- [Task-Aware 3D Affordance Segmentation via 2D Guidance and Geometric Refinement](https://mlsanity.com/paper/2511.11702/) (2025)

## Abstract (arXiv)

Functional affordance grounding requires more than recognizing an object: an agent must localize the specific region that supports an interaction, such as the handle to pull or the button to press. This is difficult for training-free vision-language pipelines because actionable regions are often small, visually ambiguous, and repeated across multiple same-category instances in a scene. We propose AFFORDMEM, a framework that grounds 3D functional affordances by remembering geometry at two levels. The first is cross-scene affordance memory: the agent maintains a category-level memory bank of RGB images with affordance regions rendered as overlays, and recalls the most informative examples at query time to guide a frozen VLM toward small operable subregions that text-only prompting consistently misses. The second is in-scene spatial memory: as the agent processes the scene, it organizes candidate instances and their 3D spatial relations into a structured scene graph, enabling the language model to resolve references over distant or currently unobserved candidates such as "the second handle from the top." AFFORDMEM requires no model fine-tuning and no target-scene annotation, using a reusable memory bank built from source scenes. On SceneFun3D, our method improves AP50 over the prior training-free state of the art by 3.23 on Split 0 and 3.7 on Split 1. Ablation studies support complementary benefits: cross-scene affordance memory improves fine-grained localization, while in-scene spatial memory provides the larger gain on spatially qualified queries. The project homepage is available at the project page.