I am testing a supervised classification algorithm and then comparing it with SVM and KNN. I tested the algorithm using Caltech-256 dataset. I wanted to know which are the commonly-used datasets like Caltech-256, which has a large in-between category similarity, making it a difficult task for classification.