1. SNU_thumbnail100.zip SNU Dataset 2201 images; thumbnail

Red rectangle - diagnosed by biopsy Black rectangle - diagnosed by clinical impression

1. quiz.zip

SNU Dataset; Quiz for AI and Dermatologist 240 images; SNU dataset QUIZ.pdf ANSWER.xlsx

doctor.xlsx doctor_AI.xlsx public.xlsx public_AI.xlsx

The exact diagnosis is listed in QUIZ_240_answer.txt

1. quiz2201.zip SNU Dataset; Quiz for AI and Dermatologist 2201 images; thumbnail (1.jpg ~ 36.jpg) quiz2201.xlsx quiz2201_AI.xlsx

answer.ods

1. disease code_with_ICD10 174 disease classes for the model training

2. SUPPLEMENTARY dataset Validation dataset that used to draw Figure S1 151 images; thumbnail Breslow thickness

Edinburgh dataset

Edinburgh dermofit image library is a commercial library.; https://licensing.edinburgh-innovations.ed.ac.uk/i/software/dermofit-image-library.html

2021.8.16

The reader test result (21 dermatologists and 23 laypersons) using the 240 images was added. - readertest240.xlsx

2022.12.15

TP / TN / FP / FN of the reader test (21 dermatologists + 26 GP + 23 laypersons) was added - readertest_statistics.xlsx

We created a test .pdf file by combining 25 images of each malignancy and 20 images of each benign conditions.Original resolution of images were extracted from the Asan test dataset and the Edinburgh Dermofit library, and doPDF (http://www.dopdf.com) was used with high resolution mode to create the test pdf. The filenames of the selected images from the Edinburgh dataset was listed in list_edin.txt. (The Edinburgh dataset is a commercial library)https://licensing.edinburgh-innovations.ed.ac.uk/item.php?item=dermofit-image-library