AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

CalliReader: Contextualizing Chinese Calligraphy via an Embedding-Aligned Vision-Language Model

Yuxuan Luo, Jiaqi Tang, Chenyi Huang, Feiyang Hao, Zhouhui Lian

ICCV 2025 · Vision-language and multimodal › Multimodal LLMs · approach: Multimodal LLM benchmarks · 6 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 e-IT embedding instruction tuning (ablation)CalliReader · CalliBench Hard tierMacro-F1: 0.390 -> 0.609Improved
add e-IT embedding instruction tuning (ablation)CalliReader · CalliBench Author VQAAcc (%): 79.31 -> 95.06Improved
add character-wise slicing and CalliAlign to InternVL2-8B (ablation)InternVL2-8B · CalliBench Hard tierMacro-F1: 0.324 -> 0.390Improved
add character-wise slicing and CalliAlign to InternVL2-8B (ablation)InternVL2-8B · CalliBench Easy tierMacro-F1: 0.617 -> 0.779Improved
use four blocks instead of two in CalliAlign (ablation)CalliAlign · label predictionAcc: 0.865 -> 0.897Improved
use layer normalization instead of global normalization (ablation)CalliAlign · label predictionAcc: 0.791 -> 0.897Improved
add ratio loss Lrat (ablation)CalliAlign · label predictionAcc: 0.897 -> 0.695Hurt
add contrastive distillation loss LCRD (ablation)CalliAlign · label predictionAcc: 0.897 -> 0.813Hurt
add slicing and CalliAlign plug-ins (ablation)CalliReader · Hard tier latencys/page: 41.29 -> 33.50Improved
add slicing and CalliAlign plug-ins (ablation)CalliReader · Easy tier latencys/page: 1.32 -> 1.48Hurt
single-task LoRA fine-tuning on page recognitionQwen2.5-VL-7B · Contextual VQA IntentAcc (%): 70.80 -> 14.43Hurt
single-task LoRA fine-tuning on page recognitionInternVL2-8B · Contextual VQA LayoutAcc (%): 45.66 -> 39.70No clear effect

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