Missed TensorFlow Dev Summit? Check out the video playlist. Watch recordings


  • Description:

The original citrus dataset contains 759 images of healthy and unhealthy citrus fruits and leaves. However, for now we only export 594 images of citrus leaves with the following labels: Black Spot, Canker, Greening, and Healthy. The exported images are in PNG format and have 256x256 pixels.

NOTE: Leaf images with Melanose label were dropped due to very small count and other non-leaf images being present in the same directory.

Dataset URL: https://data.mendeley.com/datasets/3f83gxmv57/2 License: http://creativecommons.org/licenses/by/4.0

Split Examples
'train' 594
  • Features:
    'image': Image(shape=(None, None, 3), dtype=tf.uint8),
    'image/filename': Text(shape=(), dtype=tf.string),
    'label': ClassLabel(shape=(), dtype=tf.int64, num_classes=4),
  title={A citrus fruits and leaves dataset for detection and classification of
citrus diseases through machine learning},
  author={Rauf, Hafiz Tayyab and Saleem, Basharat Ali and Lali, M Ikram Ullah
and Khan, Muhammad Attique and Sharif, Muhammad and Bukhari, Syed Ahmad Chan},
  journal={Data in brief},