uc_merced

Description:

UC Merced is a 21 class land use remote sensing image dataset, with 100 images per class. The images were manually extracted from large images from the USGS National Map Urban Area Imagery collection for various urban areas around the country. The pixel resolution of this public domain imagery is 0.3 m. Each image measures 256x256 pixels.

Homepage: http://weegee.vision.ucmerced.edu/datasets/landuse.html
Source code: tfds.image_classification.uc_merced.UcMerced
Versions:
- 2.0.0 (default): New split API (https://tensorflow.org/datasets/splits)
Download size: 317.07 MiB
Dataset size: Unknown size
Auto-cached (documentation): No
Splits:

Split	Examples
'train'	2,100

Features:

FeaturesDict({
    'filename': Text(shape=(), dtype=tf.string),
    'image': Image(shape=(None, None, 3), dtype=tf.uint8),
    'label': ClassLabel(shape=(), dtype=tf.int64, num_classes=21),
})

Supervised keys (See as_supervised doc): ('image', 'label')
Citation:

@InProceedings{Nilsback08,
   author = "Yang, Yi and Newsam, Shawn",
   title = "Bag-Of-Visual-Words and Spatial Extensions for Land-Use Classification",
   booktitle = "ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems (ACM GIS)",
   year = "2010",
}

Visualization (tfds.show_examples):

Visualization