Model Dermatology

Classification Model

Build 2017 : https://dx.medicalphoto.org; model weights
Build 2018 : https://jid2020.modelderm.com
Build 2019 : https://b2019.modelderm.com
Build 2020 : https://b2020.modelderm.com
Build 2021 : [IDerma] https://app.skindx.net

Test = Edinburgh 1300 images, 10 tumors
Malignancy = AUC value for determining malignancy
Top-1/Top-5 = Top-1 and Top-5 accuracy

Detection Model

Build 2019 : https://rcnn1.modelderm.com
Build 2020 : https://rcnn2.modelderm.com

MedicalPhoto

Since 2003 : https://medicalphoto.org

* All algorithms have been trained and maintained by Dr. Han Seung Seog [Linkdin] since 2016.