SNU and SNU MELANOMA dataset + Quiz

<b>1. SNU_thumbnail100.zip</b>SNU Dataset2201 images; thumbnail<br>Red rectangle - diagnosed by biopsyBlack rectangle - diagnosed by clinical impression<br><b>2. SNU_melanoma_thumbnail100.zip</b>SNU MELANOMA Dataset151 images; thumbnail<br><b>3. SNU_melanoma_Breslow.zip</b>SNU Dataset49 images; thumbnailSNU_melanoma_Breslow.xlsx : 1) Breslow thickness 2) invasive vs in situ 3) Age/Sex<br><br><b>4. quiz.zip</b>SNU Dataset; Quiz for AI and Dermatologist240 images; SNU datasetQUIZ.pdfANSWER.xlsxdoctor.xlsxdoctor_AI.xlsxpublic.xlsxpublic_AI.xlsx<br><b>5. quiz2201.zip</b>SNU Dataset; Quiz for AI and Dermatologist2201 images; thumbnail (1.jpg ~ 36.jpg)quiz2201.xlsxquiz2201_AI.xlsx<br>answer.ods<br><b>6. Raw output - Miss rate</b>We calculated miss rate (1-sensitivity) from the raw outputs of CNN with the whole SNU dataset (2201 images).<br>The X-axis is raw_output and Y-axis is miss rate, and the graph is close to a straight line.We used miss rate instead of raw output. Thus, we make it possible to follow up the obtained value even if the algorithm will be changed.<br><b>7. disease code_with_ICD10</b>The name of 174 training classes with ICD-10We tested with 134 disorders (Table 3 of the paper).<br><br><b># Edinburgh dataset</b>Edinburgh dermofit image library is a commercial library.; https://licensing.edinburgh-innovations.ed.ac.uk/i/software/dermofit-image-library.html<br>