What type of man against machine?

Reply to the letter to the editor ‘What type of man against machine?’ by H. Smith

“…Thirteen (22.4%) of the 58 participants showed an outperformance of the CNN in study level-I (diagnostic classification of 2A of the original publication [1]). Table 1 presents the outperformers' characteristics in comparison with the overall group of participants.…”

“…We appreciate the request [1] for a further characterisation of the group of study participants in the challenge 'Man against machine: diagnostic performance of a deep learning convolutional neural network (CNN) for dermoscopic melanoma recognition in comparison to 58 dermatologists' [2].…”

“…Continued Outperformers were defined as individual dermatologists showing sensitivity/specificity coordinates in study level-I lying above the CNN ROC curve (for reference please view Figure2Aof the original publication[1]). …”