4 ms·
I wonder if the results would be similar were the dermatologists to see the actual patient, in person, and then diagnose. And, then a photo was taken to be diag
by yumraj 8y ago
I wonder if the results would be similar were the dermatologists to see the actual patient, in person, and then diagnose. And, then a photo was taken to be diagnosed by the CNN.
In other words, while dermatologists may have been outperformed by deep learning in image classification, it is not evident if deep learning could do the same against dermatologist diagnosing in person.
Also, not clear what the overall ratio of false negative/positives was in each case.
Also, unless I missed it in the paper, I'd be curious to learn if the cases, in Fig. 4, where majority of humans and CNN disagreed, were the same where majority of humans disagreed too or not.
- nknealk 8y agoThis is extremely relevant. Factors like whether the patient has freckles or red hair often drive doctors to resect anything out of an abundance of caution. It doesn't look like these additional demographic factors were made available to clinicians. That said, the results here are still impressive. Source: I worked in clinical melanoma research for 2 years.
- ModernMech 8y agoMaybe? It's hard to say. I had a question about a mole once and what they did was take a photo of it with a special camera apparatus and they sent the image off to be diagnosed while I waited around. The doctor looked at it personally but it seems like the actual diagnosis was made by someone who never even saw me.
- sgt101 8y agoSame for me, the doctor I saw did say (after 3 seconds) "it's not a cancer but we'll take a picture anyway to be sure." I wonder if the same sort of thing would have happened if there was any real question in his mind?