esca_dataset_mixed

The zip file of esca_dataset_mixed contains a dataset with a total of 35676 images divided into 34608 inside the train folder, 534 inside the validation folder, and 534 inside the test folder. The 34608 training images are obtained by applying augmentation to a kernel of 2472 images. The esca_dataset_mixed is created by adding to the images of the "ESCA-dataset" (Mendeley Data, V1, doi: 10.17632/89cnxc58kj.1) another set of artificial images created by swapping backgrounds across the classes in which "ESCA-dataset" is partitioned.