Aesthetic Assessment of Chinese Handwritings Based on Vision Language Models

2026-03-24

AI Summary

This study addresses a critical limitation in existing Chinese handwriting automatic scoring systems, which typically provide only numerical scores without actionable feedback to guide learners’ improvement. To bridge this gap, the work proposes the first integration of Vision-Language Models (VLMs) into Chinese handwritten character aesthetic assessment. By leveraging Low-Rank Adaptation (LoRA) fine-tuning and in-context learning strategies, the model is infused with domain-specific evaluation knowledge, enabling a shift from single-score outputs to multi-level, interpretable feedback. Evaluated on multiple tracks of the CCL 2025 Chinese Handwriting Quality Assessment benchmark, the proposed approach achieves state-of-the-art performance, significantly enhancing both the practical utility of the feedback and the model’s interpretability.

Abstract

The handwriting of Chinese characters is a fundamental aspect of learning the Chinese language. Previous automated assessment methods often framed scoring as a regression problem. However, this score-only feedback lacks actionable guidance, which limits its effectiveness in helping learners improve their handwriting skills. In this paper, we leverage vision-language models (VLMs) to analyze the quality of handwritten Chinese characters and generate multi-level feedback. Specifically, we investigate two feedback generation tasks: simple grade feedback (Task 1) and enriched, descriptive feedback (Task 2). We explore both low-rank adaptation (LoRA)-based fine-tuning strategies and in-context learning methods to integrate aesthetic assessment knowledge into VLMs. Experimental results show that our approach achieves state-of-the-art performances across multiple evaluation tracks in the CCL 2025 workshop on evaluation of handwritten Chinese character quality.

Problem

• Chinese handwriting
• aesthetic assessment
• automated feedback
• handwriting evaluation
• educational technology

Innovation

• vision-language models
• aesthetic assessment
• handwritten Chinese characters
• LoRA fine-tuning
• in-context learning