Our Primary Models
Newest feature: Anonymous Filtering Support
A key new feature for autoderm_v2_2 is support for filtering anonymous images via a boolean parameter in the API call request. This allows users to flag whether the input image contains personally identifiable visual information (e.g. 2 eyes).
This feature empowers privacy-sensitive deployments—perfect for clinical settings, research, or public-facing apps requiring anonymization controls.
The Autoderm API is powered by a diverse set of models ranging from disease identification, body part location, statistical confidence and soon more use cases. We keep getting better and faster, expanding into worldwide dermatological cases not yet explored and contributing into state-of-the-art machine learning research and AI explainability.
autoderm_v2_2
Covers an extensive, more precise list of 73 skin diseases with enhanced accuracy and inference time. Stochastic results guaranteed with clear probabilistic distribution of the top-5 most likely conditions.
Newer, accurate and blazing fast
autoderm_v2_1
Covers 68 skin diseases ranked in order of the top-5 most likely cases.
autoderm_v2_0
Covers an extensive list of skin 68 diseases with enhanced accuracy for image recognition.
autoderm_v1_0
Covers an extensive list of skin 56 diseases with enhanced accuracy for image recognition.
The model can be queried from our API inference endpoint:
https://autoderm.ai/v1/query
If the model version is not provided, it will default with the latest update one being 'autoderm_v2_2'. Consult our API documentation for what parameters to use when making the requests.
Interested in specialized models or research? Reach out to us, we are hungry to accelerate!