Classification of primary specimen labels on herbarium specimen sheets dataset

This dataset contains 3,152 cropped images of primary specimen labels (also known as 'institutional labels') on herbarium specimen sheets from the University of Melbourne's herbarium.

Each image has a corresponding row in the `institutional-label-classifier-dataset.csv` file. The CSV file has three columns. The first is `path` which is the path to the image.

The second is `tag` which also has one of the following classifications

• typewriter
• printed
• handwritten
• combination
• empty

The final column is `validation` which is 1 if the image is in the validation partition and 0 if it is in the training partition.

There are 2,521 training images and 631 validation images.

For more information, see https://github.com/rbturnbull/hespi