Research map / Detection and segmentation / Open-set and few-shot detection
LMM-Det: Make Large Multimodal Models Excel in Object Detection
Jincheng Li, Chunyu Xie, Ji Ao, Dawei Leng, Yuhui Yin
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 COCO detection data to SFT (ablation) | LLaVA-7B · COCO val | AP: 0.2 -> 14.5 | Improved |
| add Object365 data at 336 resolution (ablation) | LLaVA-7B · COCO val | AP: 16.0 -> 15.6 | No clear effect |
| raise resolution 336 to 644 and add Object365 (ablation) | LLaVA-7B · COCO val | AP: 15.6 -> 38.7 | Improved |
| raise resolution 336 to 644 with COCO only (ablation) | LLaVA-7B · COCO val | AP: 16.0 -> 17.7 | Improved |
| replace CLIP-ViT with OWLv2-ViT (ablation) | LMM-Det · COCO val | AP: 38.7 -> 42.1 | Improved |
| add data distribution adjustment (ablation) | LMM-Det with OWLv2-ViT · COCO val | AP: 42.1 -> 44.2 | Improved |
| add inference optimization (per-category multi-step prediction) (ablation) | LMM-Det · COCO val | AP: 44.2 -> 47.5 | Improved |
| add inference optimization (ablation) | LMM-Det · COCO val | AR@100: 56.0 -> 63.6 | Improved |
| use OWLv2-ViT-L instead of DINOv2-L or CLIP-ViT-L (ablation) | LMM-Det (COCO only) · COCO val | AP: 16.0 -> 32.6 | Improved |
| add extra vocabularies for coordinate tokens (ablation) | LMM-Det · COCO val | AP: 32.6 -> 29.2 | Hurt |
| use beam search beam=2 instead of greedy decoding (ablation) | LMM-Det · COCO val | AP: 32.6 -> 33.0 | Mixed |
| use top-p sampling instead of greedy decoding (ablation) | LMM-Det · COCO val | AP: 32.6 -> 27.6 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 4 of 5 on most shared results.
- Salience-DETR — Salience DETR: Enhancing Detection Transformer with Hierarchical Salience Filtering Refinement (CVPR 2024) · better on 0 of 1 shared result
- Griffon-13B — Griffon: Spelling out All Object Locations at Any Granularity with Large Language Models (ECCV 2024) · better on 1 of 1 shared result
- Deformable-DETR — Deformable DETR: Deformable Transformers for End-to-End Object Detection (ICLR 2021) · better on 1 of 1 shared result
- Cascade R-CNN — Cascade R-CNN: Delving Into High Quality Object Detection (CVPR 2018) · better on 1 of 1 shared result
- Faster R-CNN — Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks (TPAMI 2016) · 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
- Ranked from Within: Ranking Large Multimodal Models Without Labels (ICML 2025 · similarity 0.86)
- ROD-MLLM: Towards More Reliable Object Detection in Multimodal Large Language Models (CVPR 2025 · similarity 0.86)
- Lumen: Unleashing Versatile Vision-Centric Capabilities of Large Multimodal Models (NeurIPS 2024 · similarity 0.85)
- On the Out-Of-Distribution Generalization of Large Multimodal Models (CVPR 2025 · similarity 0.85)
- A Concept-Based Explainability Framework for Large Multimodal Models (NeurIPS 2024 · similarity 0.85)
- MA-LMM: Memory-Augmented Large Multimodal Model for Long-Term Video Understanding (CVPR 2024 · similarity 0.84)
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.
