Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2408.13499 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909425034330112 |
|---|---|
| author | Li, Yixuan Wang, Zan Liang, Wei |
| author_facet | Li, Yixuan Wang, Zan Liang, Wei |
| contents | We propose Reasoning to Ground (R2G), a neural symbolic model that grounds the target objects within 3D scenes in a reasoning manner. In contrast to prior works, R2G explicitly models the 3D scene with a semantic concept-based scene graph; recurrently simulates the attention transferring across object entities; thus makes the process of grounding the target objects with the highest probability interpretable. Specifically, we respectively embed multiple object properties within the graph nodes and spatial relations among entities within the edges, utilizing a predefined semantic vocabulary. To guide attention transferring, we employ learning or prompting-based methods to analyze the referential utterance and convert it into reasoning instructions within the same semantic space. In each reasoning round, R2G either (1) merges current attention distribution with the similarity between the instruction and embedded entity properties or (2) shifts the attention across the scene graph based on the similarity between the instruction and embedded spatial relations. The experiments on Sr3D/Nr3D benchmarks show that R2G achieves a comparable result with the prior works while maintaining improved interpretability, breaking a new path for 3D language grounding. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2408_13499 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | R2G: Reasoning to Ground in 3D Scenes Li, Yixuan Wang, Zan Liang, Wei Computer Vision and Pattern Recognition We propose Reasoning to Ground (R2G), a neural symbolic model that grounds the target objects within 3D scenes in a reasoning manner. In contrast to prior works, R2G explicitly models the 3D scene with a semantic concept-based scene graph; recurrently simulates the attention transferring across object entities; thus makes the process of grounding the target objects with the highest probability interpretable. Specifically, we respectively embed multiple object properties within the graph nodes and spatial relations among entities within the edges, utilizing a predefined semantic vocabulary. To guide attention transferring, we employ learning or prompting-based methods to analyze the referential utterance and convert it into reasoning instructions within the same semantic space. In each reasoning round, R2G either (1) merges current attention distribution with the similarity between the instruction and embedded entity properties or (2) shifts the attention across the scene graph based on the similarity between the instruction and embedded spatial relations. The experiments on Sr3D/Nr3D benchmarks show that R2G achieves a comparable result with the prior works while maintaining improved interpretability, breaking a new path for 3D language grounding. |
| title | R2G: Reasoning to Ground in 3D Scenes |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2408.13499 |