Dermatology Machine Learning!

Learn more

About DermML

This application combines textual and image feature extraction based on Machine Learning. If you are a Dermatologist, please join us in training and testing this application. UpVote / Downvote based on the relevance of the rule in the diagnosis of the given condition. Registered users can delete irrelevant rules and train the feature extractor based on your own clinical images. When you train the feature extractor, please select the lesion. The selected part is uploaded, but not saved on the server. Only the feature descriptors are saved. Feature descriptors are a set of numbers, and the original image cannot be recreated from these numbers. Contact Bell Eapen [admin (at) gulfdoctor.net] for an account.

Please watch the video below:

This application is still in the alpha stage and is being tested. Please send your feedbacks to admin (at) gulfdoctor.net. This application is for medical professionals in resource-deprived areas to get a list of dermatological differential diagnosis for a given patient. The algorithm offers rare differentials first as they are the ones likely to be missed. The accuracy improves with crowdsourced curating of rules with the help of domain experts like you.

Publications related to this project:

  1. Eapen, Bell Raj. "ONTODerm-A domain ontology for dermatology." Dermatology online journal 14.6 (2008)
  2. Sáenz, J. P., Novoa, M. P., Correal, D., & Eapen, B. R. (2018). On Using a Mobile Application to Support Teledermatology: A Case Study in an Underprivileged Area in Colombia. International journal of telemedicine and applications, 2018.
  3. Sáenz, J. P., Novoa, M. P., Correal, D., & Eapen, B. R. (2016, November). Skinhealth, A Mobile Application for Supporting Teledermatology: A Case Study in a Rural Area in Colombia. In International Conference on Wireless Mobile Communication and Healthcare (pp. 160-163). Springer, Cham.

 

People:

Bell Eapen

Bell Eapen