Qualification of a dermascope imaging device

A dermascope imaging device comprising:

• a processor and storage, wherein the processor is configured to perform qualifying the dermascope imaging device for use with an image classification algorithm, said qualifying comprising:
• accessing a qualification data set comprising a plurality of pairs of images of skin lesions, wherein each pair comprises an image of a skin lesion captured by an unqualified dermascope imaging device and an image of the skin lesion captured by a qualified dermascope imaging device;
• computing, using the image classification algorithm, a confidence value of classification of each image, wherein the image classification algorithm is a machine-learning based image classification algorithm;
• measuring a similarity metric between the unqualified and qualified dermascope imaging device using differences in the confidence values between images of each pair;
• and
• qualifying the unqualified dermascope imaging device for use with the image classification algorithm in response a comparison between the similarity metric and a similarity threshold, wherein the processor is further configured to calculate the similarity threshold, performing:
• obtaining a calibration data set of skin lesion images where a confidence value of each image computed using the image classification algorithm is known and wherein each skin lesion image is labelled as a lesion type comprising any of:
• malignant, premalignant, benign;
• applying a same series of parameterised distortions and/or transformations to each image from the calibration data set;
• computing, using the image classification algorithm, a confidence value of a classification of each image of the distorted and/or transformed images;
• measuring the similarity metric between each image of the calibration data set and each corresponding distorted/transformed image;
• determining a sensitivity metric for each skin lesion image of type malignant and premalignant, the sensitivity metric having a confidence interval;
• recording the similarity metric and sensitivity metric for each pair comprising a calibration image and a distorted/transformed image;
• and
• selecting a largest similarity metric before any similarity metric drops below its confidence interval as the similarity threshold.