Yucatan pollen grain dataset

For convenience we've added a script to divide the dataset into training, validation and test subsets run the divideSets.sh script, for example:

./divideSets.sh pathToImages pathTo/obj.names 80 10
  • The first argument is the path to the images.
  • The second argument is the path to the obj.names file, which contains the name of the classes.
  • The third argument is the percentage of each class to use for the training set.
  • The fourth argument is the percentage of each class to use for the validation set (the rest is used for the test set).

a

Top left: example of a Box catzín image; top right: example of a Tajonal image; bottom left: example of a Jabín image; bottom right: example of a Chaká image.