Skin Disease Classification with LVLMs: An Empirical Study

Xinyi Zeng; Zimu Wang; Haiyang Zhang; Yiming Luo; Wei Wang

doi:10.1109/CSCWD64889.2025.11033252

Skin Disease Classification with LVLMs: An Empirical Study

Xinyi Zeng, Zimu Wang, Haiyang Zhang^*, Yiming Luo, Wei Wang

^*Corresponding author for this work

Research output: Chapter in Book or Report/Conference proceeding › Conference Proceeding › peer-review

Abstract

Skin diseases pose significant challenges to accurate and efficient diagnosis, often due to their diverse and complex representations. This study investigates the capabilities and limitations of Large Vision-Language Models (LVLMs) in addressing these challenges through skin disease classification tasks. We evaluated LVLMs in zero-shot, few-shot, and finetuning scenarios, exploring their performance, bias, and potential for improvement. Results show that LVLMs lack perceptual granularity in skin disease, though positive signals are also observed. Our findings underscore the necessity for domain- specific optimisation and highlight opportunities for advancing LVLMs in medical diagnostics through innovative strategies and collaborative efforts.

Original language	English
Title of host publication	Proceedings of the 2025 28th International Conference on Computer Supported Cooperative Work in Design, CSCWD 2025
Editors	Weiming Shen, Weiming Shen, Marie-Helene Abel, Nada Matta, Jean-Paul Barthes, Junzhou Luo, Jinghui Zhang, Haibin Zhu, Kunkun Peng
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	2304-2309
Number of pages	6
Edition	2025
ISBN (Electronic)	9798331513054
DOIs	https://doi.org/10.1109/CSCWD64889.2025.11033252
Publication status	Published - 2025
Event	28th International Conference on Computer Supported Cooperative Work in Design, CSCWD 2025 - Compiegne, France Duration: 5 May 2025 → 7 May 2025

Conference

Conference	28th International Conference on Computer Supported Cooperative Work in Design, CSCWD 2025
Country/Territory	France
City	Compiegne
Period	5/05/25 → 7/05/25

Keywords

large vision-language models
medical intelligent diagnostics
skin disease classification

Access to Document

10.1109/CSCWD64889.2025.11033252

Cite this

Zeng, X., Wang, Z., Zhang, H., Luo, Y., & Wang, W. (2025). Skin Disease Classification with LVLMs: An Empirical Study. In W. Shen, W. Shen, M.-H. Abel, N. Matta, J.-P. Barthes, J. Luo, J. Zhang, H. Zhu, & K. Peng (Eds.), Proceedings of the 2025 28th International Conference on Computer Supported Cooperative Work in Design, CSCWD 2025 (2025 ed., pp. 2304-2309). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/CSCWD64889.2025.11033252

Zeng, Xinyi ; Wang, Zimu ; Zhang, Haiyang et al. / Skin Disease Classification with LVLMs : An Empirical Study. Proceedings of the 2025 28th International Conference on Computer Supported Cooperative Work in Design, CSCWD 2025. editor / Weiming Shen ; Weiming Shen ; Marie-Helene Abel ; Nada Matta ; Jean-Paul Barthes ; Junzhou Luo ; Jinghui Zhang ; Haibin Zhu ; Kunkun Peng. 2025. ed. Institute of Electrical and Electronics Engineers Inc., 2025. pp. 2304-2309

@inproceedings{1d44a2db9fb748a5adb2d85612920a5f,

title = "Skin Disease Classification with LVLMs: An Empirical Study",

abstract = "Skin diseases pose significant challenges to accurate and efficient diagnosis, often due to their diverse and complex representations. This study investigates the capabilities and limitations of Large Vision-Language Models (LVLMs) in addressing these challenges through skin disease classification tasks. We evaluated LVLMs in zero-shot, few-shot, and finetuning scenarios, exploring their performance, bias, and potential for improvement. Results show that LVLMs lack perceptual granularity in skin disease, though positive signals are also observed. Our findings underscore the necessity for domain- specific optimisation and highlight opportunities for advancing LVLMs in medical diagnostics through innovative strategies and collaborative efforts.",

keywords = "large vision-language models, medical intelligent diagnostics, skin disease classification",

author = "Xinyi Zeng and Zimu Wang and Haiyang Zhang and Yiming Luo and Wei Wang",

note = "Publisher Copyright: {\textcopyright} 2025 IEEE.; 28th International Conference on Computer Supported Cooperative Work in Design, CSCWD 2025 ; Conference date: 05-05-2025 Through 07-05-2025",

year = "2025",

doi = "10.1109/CSCWD64889.2025.11033252",

language = "English",

pages = "2304--2309",

editor = "Weiming Shen and Weiming Shen and Marie-Helene Abel and Nada Matta and Jean-Paul Barthes and Junzhou Luo and Jinghui Zhang and Haibin Zhu and Kunkun Peng",

booktitle = "Proceedings of the 2025 28th International Conference on Computer Supported Cooperative Work in Design, CSCWD 2025",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

edition = "2025",

}

Zeng, X, Wang, Z, Zhang, H, Luo, Y & Wang, W 2025, Skin Disease Classification with LVLMs: An Empirical Study. in W Shen, W Shen, M-H Abel, N Matta, J-P Barthes, J Luo, J Zhang, H Zhu & K Peng (eds), Proceedings of the 2025 28th International Conference on Computer Supported Cooperative Work in Design, CSCWD 2025. 2025 edn, Institute of Electrical and Electronics Engineers Inc., pp. 2304-2309, 28th International Conference on Computer Supported Cooperative Work in Design, CSCWD 2025, Compiegne, France, 5/05/25. https://doi.org/10.1109/CSCWD64889.2025.11033252

Skin Disease Classification with LVLMs: An Empirical Study. / Zeng, Xinyi; Wang, Zimu; Zhang, Haiyang et al.
Proceedings of the 2025 28th International Conference on Computer Supported Cooperative Work in Design, CSCWD 2025. ed. / Weiming Shen; Weiming Shen; Marie-Helene Abel; Nada Matta; Jean-Paul Barthes; Junzhou Luo; Jinghui Zhang; Haibin Zhu; Kunkun Peng. 2025. ed. Institute of Electrical and Electronics Engineers Inc., 2025. p. 2304-2309.

Research output: Chapter in Book or Report/Conference proceeding › Conference Proceeding › peer-review

TY - GEN

T1 - Skin Disease Classification with LVLMs

T2 - 28th International Conference on Computer Supported Cooperative Work in Design, CSCWD 2025

AU - Zeng, Xinyi

AU - Wang, Zimu

AU - Zhang, Haiyang

AU - Luo, Yiming

AU - Wang, Wei

PY - 2025

Y1 - 2025

N2 - Skin diseases pose significant challenges to accurate and efficient diagnosis, often due to their diverse and complex representations. This study investigates the capabilities and limitations of Large Vision-Language Models (LVLMs) in addressing these challenges through skin disease classification tasks. We evaluated LVLMs in zero-shot, few-shot, and finetuning scenarios, exploring their performance, bias, and potential for improvement. Results show that LVLMs lack perceptual granularity in skin disease, though positive signals are also observed. Our findings underscore the necessity for domain- specific optimisation and highlight opportunities for advancing LVLMs in medical diagnostics through innovative strategies and collaborative efforts.

AB - Skin diseases pose significant challenges to accurate and efficient diagnosis, often due to their diverse and complex representations. This study investigates the capabilities and limitations of Large Vision-Language Models (LVLMs) in addressing these challenges through skin disease classification tasks. We evaluated LVLMs in zero-shot, few-shot, and finetuning scenarios, exploring their performance, bias, and potential for improvement. Results show that LVLMs lack perceptual granularity in skin disease, though positive signals are also observed. Our findings underscore the necessity for domain- specific optimisation and highlight opportunities for advancing LVLMs in medical diagnostics through innovative strategies and collaborative efforts.

KW - large vision-language models

KW - medical intelligent diagnostics

KW - skin disease classification

UR - https://www.scopus.com/pages/publications/105011516310

U2 - 10.1109/CSCWD64889.2025.11033252

DO - 10.1109/CSCWD64889.2025.11033252

M3 - Conference Proceeding

AN - SCOPUS:105011516310

SP - 2304

EP - 2309

BT - Proceedings of the 2025 28th International Conference on Computer Supported Cooperative Work in Design, CSCWD 2025

A2 - Shen, Weiming

A2 - Abel, Marie-Helene

A2 - Matta, Nada

A2 - Barthes, Jean-Paul

A2 - Luo, Junzhou

A2 - Zhang, Jinghui

A2 - Zhu, Haibin

A2 - Peng, Kunkun

PB - Institute of Electrical and Electronics Engineers Inc.

Y2 - 5 May 2025 through 7 May 2025

ER -

Zeng X, Wang Z, Zhang H, Luo Y, Wang W. Skin Disease Classification with LVLMs: An Empirical Study. In Shen W, Shen W, Abel MH, Matta N, Barthes JP, Luo J, Zhang J, Zhu H, Peng K, editors, Proceedings of the 2025 28th International Conference on Computer Supported Cooperative Work in Design, CSCWD 2025. 2025 ed. Institute of Electrical and Electronics Engineers Inc. 2025. p. 2304-2309 doi: 10.1109/CSCWD64889.2025.11033252

Skin Disease Classification with LVLMs: An Empirical Study

Abstract

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this