We propose that the GUI allow/force the operator to assign a "sample type" to each image. This will be a simple string chosen from a fixed, registered set of strings. Each sample type will offer a fixed, registered set of classifiers that are allowed to operate on that sample type. The operator should select what classifiers they want applied to the image.
Eventually, the backend should apply all of the N selected classifiers to the images and produce N reports which are added to the case and visible to the user.
@Nipul12 , I am assign to you, but you may wish to reassign to @sanchaydevnath or @avinashbaskaran