This dataset consists of 4502 images of healthy and unhealthy plant leaves divided into 22 categories by species and state of health. The images are in high resolution JPG format.

There are no files with label prefix 0000, therefore label encoding is shifted by one (e.g. file with label prefix 0001 gets encoded label 0).

Note: Each image is a separate download. Some might rarely fail, therefore make sure to restart if that happens. An exception will be raised in case one of the downloads repeatedly fails.

Dataset URL: https://data.mendeley.com/datasets/hb74ynkjcn/1 License: http://creativecommons.org/licenses/by/4.0

To use this dataset:

import tensorflow_datasets as tfds

ds = tfds.load('plant_leaves', split='train')
for ex in ds.take(4):
 print(ex)

See the guide for more informations on tensorflow_datasets.

https://storage.googleapis.com/tfds-data/visualization/fig/plant_leaves-0.1.0.png" alt="Visualization" width="500px">