i started my project for skin lesin classification … i initially used tenorflow and the dermnet dataset but i didnt edit augmented the images , cus i didnt know how to what to my text book was limited , can someone please suggest some precedure to make a skin lesion detection/classification im using isic 2019 and dermnet datasets and scin dataset as well isic is on dermascopic images but the sccin and dermnet are full body regions … can you suggest how to make something good , or are there any publically available projects … i thing my knowledge of augmentations , preprocesses is limited i have athe books but i dont know the perfect technique i already tried transferlearning accuracy is bad .