AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

Leveraging the Power of MLLMs for Gloss-Free Sign Language Translation

Jungeun Kim, Hyeongwoo Jeon, Jongseong Bae, Ha Young Kim

ICCV 2025 · Vision-language and multimodal › Multimodal LLMs · approach: Benchmarks and analysis of MLLMs · 19 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
add MLLM descriptions to images without alignment (ablation)MMSLT · PHOENIX14T testBLEU-4: 13.19 -> 12.78Hurt
add multimodal-language contrastive alignment (ablation)MMSLT · PHOENIX14T testBLEU-4: 12.78 -> 25.04Improved
add GSD-MLLM descriptions on top of alignment (ablation)MMSLT · PHOENIX14T testBLEU-4: 23.77 -> 25.04Improved
add description mapper on top of descriptions + alignment (ablation)MMSLT · PHOENIX14T testBLEU-4: 25.04 -> 25.73Improved
add description mapper without alignment (ablation)images + descriptions · PHOENIX14T testBLEU-4: 12.78 -> 24.21Improved
use LLaVA-OneVision 7B for descriptions instead of InternVL2 8B (ablation)MMSLT · PHOENIX14T testBLEU-4: 23.16 -> 25.73Improved
use detailed prompt (3) instead of simple prompt (2) (ablation)GSD-MLLM · PHOENIX14T testBLEU-4: 25.00 -> 25.73Improved
use in-context prompt (5) instead of detailed prompt (3) (ablation)GSD-MLLM · PHOENIX14T testBLEU-4: 25.73 -> 23.10Hurt
use BERT instead of Qwen2 as description encoder (ablation)MMSLT · PHOENIX14T testBLEU-4: 22.64 -> 25.73Improved
use BERT instead of SentenceBERT as description encoder (ablation)MMSLT · PHOENIX14T testBLEU-4: 25.40 -> 25.73Improved
use MLLM descriptions alone without visual features (ablation)MMSLT · PHOENIX14T testBLEU-4: 8.99 vs 25.73 with visualHurt
MMSLT vs previous best gloss-free (SignLLM) on CSL-DailyMMSLT · CSL-Daily testBLEU-4: 15.75 -> 21.11Improved

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