Algorithm 2: Diagnosis_Service

Input: skin_image Output: P[p0, …,pC] ∈ R   //C is the number of skin disease classes

1   Function: get_diagnosis (skin_image) 2    Init: size ← model input dimension, std ← model normalization factor      //Image pre-processing 3    skin_image ← skin_image.rsize(size, size)//Resize the image to size x size 4    img_array[size, size] ← convert_to_array(skin_image)//Convert image to an array      //Normalization 5    For x in img_array 6       For y in img_array[x] 7          img_array[x, y] ← img_array[x, y]/std 8       End For 9    End For      //Classification 10    model ← load_model() //load trained model 11    P ← model.predict(img_array) 12    Return P