The algorithm would then discover this labeled selection of photographs to differentiate the styles and their characteristics, for instance circles with no corners and squares with 4 equivalent sides. Right after training about the dataset of images, the technique can see a new picture and identify what shape it finds. The amount and complexity of