AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

How Can Objects Help Video-Language Understanding?

Zitian Tang, Shijie Wang, Junho Cho, Jaewook Yoo, Chen Sun

ICCV 2025 · Vision-language and multimodal › Multimodal LLMs · approach: VLM capability analysis · 4 citations (OpenAlex)

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 changedWhereMeasuredEffect
represent boxes as quantized plain text instead of a learned embedding projector (ablation)ObjectMLLM · CLEVRER-MCaccuracy: 64.9 -> 77.6Improved
represent boxes as plain text instead of embedding projector, with 10k training samples (ablation)ObjectMLLM · CLEVRER-MC (10k samples)accuracy: 44.5 -> 63.8Improved
zero-initialize the box embedding projector instead of random init (ablation)Embedding projector · CLEVRER-MCaccuracy: 42.8 -> 64.9Improved
use MLP with more layers for box embedding projector (ablation)Embedding projector · CLEVRER-MCaccuracy: 64.9 -> 65.0No clear effect
use 9 tokens per box in embedding projector instead of 1 (ablation)Embedding projector · CLEVRER-MCaccuracy: 64.9 -> 63.4Hurt
add object bounding boxes to captions (ablation)ObjectMLLM (LLaMA3) · STARaccuracy: 60.1 -> 64.4Improved
add object bounding boxes to captions (ablation)ObjectMLLM (LLaMA3) · CLEVRER-MCaccuracy: 47.8 -> 75.5Improved
add object bounding boxes to captions (ablation)ObjectMLLM (LLaMA3) · NExT-QA and IntentQAaccuracy: 76.6 -> 76.6 (NExT-QA); 75.7 -> 75.6 (IntentQA)No clear effect
add CLIP visual embeddings on top of caption and box (ablation)ObjectMLLM (LLaMA3) · CLEVRER-MCaccuracy: 75.5 -> 75.4No clear effect
add boxes to pre-trained VideoLLaMA2 and LoRA fine-tuneVideoLLaMA2-7B · CLEVRER-MCaccuracy: 67.9 -> 77.6Improved
add boxes to VideoLLaMA2 without fine-tuning (zero-shot)VideoLLaMA2-7B · CLEVRER-MCaccuracy: 45.6 -> 34.4Hurt
train modality-by-modality instead of jointly on all modalities (ablation)ObjectMLLM (LLaMA3) · CLEVRER-MCaccuracy: 29.5 -> 75.4Improved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy