Research map / Vision-language and multimodal / Multimodal LLMs
GeoProg3D: Compositional Visual Reasoning for City-Scale 3D Language Fields
Shunsuke Yasuki, Taiki Miyanishi, Nakamasa Inoue, Shuhei Kurita, Koya Sakamoto, Daichi Azuma, Masato Taki, Yutaka Matsuo
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add GV-APIs and visual programming on top of the GCLF language field | GCLF · GeoEval3D GoogleEarth GRD | localization accuracy (%): 20.09 -> 45.20 | Improved |
| add GV-APIs and visual programming on top of the GCLF language field | GCLF · GeoEval3D UrbanScene3D GRD | localization accuracy (%): 6.98 -> 30.23 | Improved |
| use GCLF (tree-structured, georeferenced language field) instead of LangSplat (ablation) | LangSplat · GeoEval3D GoogleEarth GRD | localization accuracy (%): 14.15 -> 20.09 | Improved |
| use GCLF instead of LangSplat on the 5.0x10^6 m^2 scene (ablation) | LangSplat · GeoEval3D UrbanScene3D | localization: OOM -> 6.98 | Improved |
| add GCLF's tree structure (more Gaussians) versus LangSplat (ablation) | GCLF · rendering speed | rendering time (ms): 2.73 -> 14.46 | Hurt |
| add GetLandmarkSeg module (ablation) | GeoProg3D · GeoEval3D GoogleEarth GRD | accuracy (%): 6.26 -> 45.20 | Improved |
| add GetLandmarkSeg module (ablation) | GeoProg3D · GeoEval3D GoogleEarth CMP | accuracy (%): 0.00 -> 59.73 | Improved |
| add SegDirection module (ablation) | GeoProg3D · GeoEval3D GoogleEarth GRD | accuracy (%): 26.01 -> 45.20 | Improved |
| add SegAround module (ablation) | GeoProg3D · GeoEval3D GoogleEarth SPR | accuracy (%): 36.76 -> 64.00 | Improved |
| add SegBetween module (ablation) | GeoProg3D · GeoEval3D GoogleEarth CMP | accuracy (%): 60.53 -> 59.73 | Mixed |
| add LargestSeg module (ablation) | GeoProg3D · GeoEval3D GoogleEarth CMP | accuracy (%): 44.74 -> 59.73 | Improved |
| add LargestSeg module (ablation) | GeoProg3D · GeoEval3D GoogleEarth GRD | accuracy (%): 45.20 -> 45.20 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- TEOChat — TEOChat: A Large Vision-Language Assistant for Temporal Earth Observation Data (ICLR 2025) · better on 2 of 2 shared results
- GeoChat — GeoChat: Grounded Large Vision-Language Model for Remote Sensing (CVPR 2024) · better on 1 of 1 shared result
- LHRS-BOT — LHRS-Bot: Empowering Remote Sensing with VGI-Enhanced Large Multimodal Language Model (ECCV 2024) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- GRE Suite: Geo-localization Inference via Fine-Tuned Vision-Language Models and Enhanced Reasoning Chains (NeurIPS 2025 · similarity 0.90)
- From Flatland to Space: Teaching Vision-Language Models to Perceive and Reason in 3D (NeurIPS 2025 · similarity 0.90)
- SpatialReasoner: Towards Explicit and Generalizable 3D Spatial Reasoning (NeurIPS 2025 · similarity 0.90)
- SpatialRGPT: Grounded Spatial Reasoning in Vision-Language Models (NeurIPS 2024 · similarity 0.89)
- Recognition through Reasoning: Reinforcing Image Geo-localization with Large Vision-Language Models (NeurIPS 2025 · similarity 0.89)
- CitySTAR: Agent-Driven Structured and Topology-Aware Reasoning for Open-Vocabulary Urban 3D Grounding (NeurIPS 2026 · similarity 0.89)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
