Research map / Vision-language and multimodal / Multimodal LLMs
Uni3DL: A Unified Model for 3D Vision-Language Understanding
Xiang Li, Jian Ding, Zhaoyang Chen, Mohamed Elhoseiny
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 multi-task pretraining before task fine-tuning (ablation) | Uni3DL · ScanNet v2 val semantic segmentation | mIoU: 72.3 -> 76.2 | Improved |
| add multi-task pretraining before task fine-tuning (ablation) | Uni3DL · Text2Shape retrieval | R@5: 7.7 -> 19.7 | Improved |
| add instance segmentation task in pretraining (ablation) | Uni3DL · ScanRefer grounded segmentation | Acc@0.25: 35.8 -> 37.8 | Improved |
| add captioning task in pretraining (ablation) | Uni3DL · Cap3D retrieval | T2S R@5: 12.8 -> 15.5 | Improved |
| add retrieval task in pretraining (ablation) | Uni3DL · Cap3D captioning | BLEU-1: 13.5 -> 16.8 | Improved |
| scale object-task weights beta from 1 up to 5 (ablation) | Uni3DL · ScanRefer / Cap3D | multiple: ScanRefer Acc@0.25 37.8 (beta 1) -> 35.2 (beta 5) | Mixed |
| use alternate training between object and scene tasks instead of joint training (ablation) | Uni3DL · ScanRefer | Acc@0.25: 37.8 -> 36.8 | Hurt |
| use unified Uni3DL instead of task-specific mask-attention-free model on S3DIS instance segmentation | Uni3DL · S3DIS Area 5 instance segmentation | mAP50: 69.1 -> 65.3 | Hurt |
Closest papers by meaning
- Unifying 2D and 3D Vision-Language Understanding (ICML 2025 · similarity 0.93)
- A Unified Framework for 3D Scene Understanding (NeurIPS 2024 · similarity 0.91)
- Unveiling the Mist over 3D Vision-Language Understanding: Object-centric Evaluation with Chain-of-Analysis (CVPR 2025 · similarity 0.91)
- Language-Image Models with 3D Understanding (ICLR 2025 · similarity 0.90)
- Lexicon3D: Probing Visual Foundation Models for Complex 3D Scene Understanding (NeurIPS 2024 · similarity 0.90)
- Uni3D: Exploring Unified 3D Representation at Scale (ICLR 2024 · similarity 0.90)
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.
