GeoDANO: Geometric VLM with Domain Agnostic Vision Encoder
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866916970547380224 |
|---|---|
| author | Cho, Seunghyuk Qin, Zhenyue Liu, Yang Choi, Youngbin Lee, Seungbeom Kim, Dongwoo |
| author_facet | Cho, Seunghyuk Qin, Zhenyue Liu, Yang Choi, Youngbin Lee, Seungbeom Kim, Dongwoo |
| contents | We introduce GeoDANO, a geometric vision-language model (VLM) with a domain-agnostic vision encoder, for solving plane geometry problems. Although VLMs have been employed for solving geometry problems, their ability to recognize geometric features remains insufficiently analyzed. To address this gap, we propose a benchmark that evaluates the recognition of visual geometric features, including primitives such as dots and lines, and relations such as orthogonality. Our preliminary study shows that vision encoders often used in general-purpose VLMs, e.g., OpenCLIP, fail to detect these features and struggle to generalize across domains. To overcome the limitation, we develop GeoCLIP, a CLIP-based model trained on synthetic geometric diagram--caption pairs. Benchmark results show that GeoCLIP outperforms existing vision encoders in recognizing geometric features. We then propose our VLM, GeoDANO, which augments GeoCLIP with a domain adaptation strategy for unseen diagram styles. GeoDANO outperforms specialized methods for plane geometry problems and GPT-4o on MathVerse. The implementation is available at https://github.com/ml-postech/GeoDANO. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2502_11360 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | GeoDANO: Geometric VLM with Domain Agnostic Vision Encoder Cho, Seunghyuk Qin, Zhenyue Liu, Yang Choi, Youngbin Lee, Seungbeom Kim, Dongwoo Computer Vision and Pattern Recognition Computation and Language We introduce GeoDANO, a geometric vision-language model (VLM) with a domain-agnostic vision encoder, for solving plane geometry problems. Although VLMs have been employed for solving geometry problems, their ability to recognize geometric features remains insufficiently analyzed. To address this gap, we propose a benchmark that evaluates the recognition of visual geometric features, including primitives such as dots and lines, and relations such as orthogonality. Our preliminary study shows that vision encoders often used in general-purpose VLMs, e.g., OpenCLIP, fail to detect these features and struggle to generalize across domains. To overcome the limitation, we develop GeoCLIP, a CLIP-based model trained on synthetic geometric diagram--caption pairs. Benchmark results show that GeoCLIP outperforms existing vision encoders in recognizing geometric features. We then propose our VLM, GeoDANO, which augments GeoCLIP with a domain adaptation strategy for unseen diagram styles. GeoDANO outperforms specialized methods for plane geometry problems and GPT-4o on MathVerse. The implementation is available at https://github.com/ml-postech/GeoDANO. |
| title | GeoDANO: Geometric VLM with Domain Agnostic Vision Encoder |
| topic | Computer Vision and Pattern Recognition Computation and Language |
| url | https://arxiv.org/abs/2502.11360 |