SketchJudge: A Diagnostic Benchmark for Grading Hand-drawn Diagrams with Multimodal Large Language Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Su, Yuhang, Wang, Mei, Zhong, Yaoyao, Li, Guozhang, Li, Shixing, Feng, Yihan, Huang, Hua
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912817278353408
author Su, Yuhang
Wang, Mei
Zhong, Yaoyao
Li, Guozhang
Li, Shixing
Feng, Yihan
Huang, Hua
author_facet Su, Yuhang
Wang, Mei
Zhong, Yaoyao
Li, Guozhang
Li, Shixing
Feng, Yihan
Huang, Hua
contents While Multimodal Large Language Models (MLLMs) have achieved remarkable progress in visual understanding, they often struggle when faced with the unstructured and ambiguous nature of human-generated sketches. This limitation is particularly pronounced in the underexplored task of visual grading, where models should not only solve a problem but also diagnose errors in hand-drawn diagrams. Such diagnostic capabilities depend on complex structural, semantic, and metacognitive reasoning. To bridge this gap, we introduce SketchJudge, a novel benchmark tailored for evaluating MLLMs as graders of hand-drawn STEM diagrams. SketchJudge encompasses 1,015 hand-drawn student responses across four domains: geometry, physics, charts, and flowcharts, featuring diverse stylistic variations and distinct error types. Evaluations on SketchJudge demonstrate that even advanced MLLMs lag significantly behind humans, validating the benchmark's effectiveness in exposing the fragility of current vision-language alignment in symbolic and noisy contexts. All data, code, and evaluation scripts are publicly available at https://github.com/yuhangsu82/SketchJudge.
format Preprint
id arxiv_https___arxiv_org_abs_2601_06944
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle SketchJudge: A Diagnostic Benchmark for Grading Hand-drawn Diagrams with Multimodal Large Language Models
Su, Yuhang
Wang, Mei
Zhong, Yaoyao
Li, Guozhang
Li, Shixing
Feng, Yihan
Huang, Hua
Computer Vision and Pattern Recognition
Artificial Intelligence
While Multimodal Large Language Models (MLLMs) have achieved remarkable progress in visual understanding, they often struggle when faced with the unstructured and ambiguous nature of human-generated sketches. This limitation is particularly pronounced in the underexplored task of visual grading, where models should not only solve a problem but also diagnose errors in hand-drawn diagrams. Such diagnostic capabilities depend on complex structural, semantic, and metacognitive reasoning. To bridge this gap, we introduce SketchJudge, a novel benchmark tailored for evaluating MLLMs as graders of hand-drawn STEM diagrams. SketchJudge encompasses 1,015 hand-drawn student responses across four domains: geometry, physics, charts, and flowcharts, featuring diverse stylistic variations and distinct error types. Evaluations on SketchJudge demonstrate that even advanced MLLMs lag significantly behind humans, validating the benchmark's effectiveness in exposing the fragility of current vision-language alignment in symbolic and noisy contexts. All data, code, and evaluation scripts are publicly available at https://github.com/yuhangsu82/SketchJudge.
title SketchJudge: A Diagnostic Benchmark for Grading Hand-drawn Diagrams with Multimodal Large Language Models
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2601.06944