Researchmultimodal llmtraditional chinese medicineclip vitllama 7b
TongueVLM Achieves Multimodal Tongue Diagnosis Accuracy
7.3
Researchers from Hefei University of Technology and collaborators developed TongueVLM, a multimodal large model for traditional Chinese medicine tongue-image diagnosis, published in JMIR Medical Informatics (2026). The LLaMA-based 7B-parameter model uses CLIP-ViT visual encoding and modal fusion, evaluated on three test datasets (3,000 samples each) and achieved 79.8%, 78.6%, and 60.7% accuracy, outperforming baseline VLMs.
Key Points
- 1Develops TongueVLM, a 7B-parameter multimodal model aligning tongue images with TCM terminology
- 2Demonstrates superior accuracy versus LLaVA-OneVision across three tasks, improving by up to 9.1%
- 3Enables automated tongue description, constitution reasoning, and potential integration into TCM diagnostic systems
Scoring Rationale
Strong peer-reviewed evaluation and clear empirical gains drive the score, but niche TCM focus limits wider impact.
Sources
Public references used for this report.
Practice with real Ad Tech data
90 SQL & Python problems · 15 industry datasets
Used by DS/ML engineers at top companies
Active Search Campaigns by BudgetEasyHigh CPC Clicks & Poor Landing PagesMediumCampaign ROAS by Attribution ModelHard
250 free problems · No credit card
See all Ad Tech problems