Cross-Dataset Testbed Image Classification Dataset

Cross-Dataset Testbed is a cross-dataset image classification dataset based on Decaf7. The dataset contains images from 3 domains, including 3,847 images from the Caltech256 dataset, 4,000 images from the ImageNet dataset, and 2,626 images from the SUN dataset. The dataset contains a total of 10,473 images from 40 categories.