Derm1M: 1 Million Dermatological Image-Text Pairs with Expert Ontology

Derm1M is a large-scale vision-language dataset for dermatology containing 1,029,761 image-text pairs from 403,563 unique images. The dataset covers 390 skin conditions organized in a four-level expert ontology and includes 130 clinical concepts. It was created by author redlessone and last updated on the Hugging Face platform in October 2025.

Use Cases

• Training vision-language models for dermatological image captioning based on the rich contextual captions.
• Developing diagnostic classifiers for 390 skin conditions based on the expert ontology.
• Researching clinical concept extraction from medical images based on the 130 included concepts.
• Benchmarking multimodal medical AI systems based on the dataset's million-scale size.

Strengths

• Large scale with 1,029,761 image-text pairs from 403,563 unique images.
• Comprehensive coverage of 390 skin conditions organized in a four-level expert ontology.
• Includes 130 clinical concepts for detailed annotation.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• License is unknown, which may restrict commercial or research use.
• Data may reflect geographic or source bias inherent to its collection method.

Provenance

Source

redlessone on Hugging Face

Collection Method

Likely aggregated from various dermatological sources to create image-text pairs.

Time Range

null

Freshness

Last updated 2025-10-23 08:43:50.

Geography

Region 'us' is indicated in platform tags, but specific spatial coverage is unknown.