• Description:

The Stanford Dogs dataset contains images of 120 breeds of dogs from around the world. This dataset has been built using images and annotation from ImageNet for the task of fine-grained image categorization. There are 20,580 images, out of which 12,000 are used for training and 8580 for testing. Class labels and bounding box annotations are provided for all the 12,000 images.

Split Examples
'test' 8,580
'train' 12,000
  • Feature structure:
    'image': Image(shape=(None, None, 3), dtype=uint8),
    'image/filename': Text(shape=(), dtype=string),
    'label': ClassLabel(shape=(), dtype=int64, num_classes=120),
    'objects': Sequence({
        'bbox': BBoxFeature(shape=(4,), dtype=float32),
  • Feature documentation:
Feature Class Shape Dtype Description
image Image (None, None, 3) uint8
image/filename Text string
label ClassLabel int64
objects Sequence
objects/bbox BBoxFeature (4,) float32


  • Citation:
author = "Aditya Khosla and Nityananda Jayadevaprakash and Bangpeng Yao and
          Li Fei-Fei",
title = "Novel Dataset for Fine-Grained Image Categorization",
booktitle = "First Workshop on Fine-Grained Visual Categorization,
             IEEE Conference on Computer Vision and Pattern Recognition",
year = "2011",
month = "June",
address = "Colorado Springs, CO",
        AUTHOR = {Deng, J. and Dong, W. and Socher, R. and Li, L.-J. and
                  Li, K. and Fei-Fei, L.},
        TITLE = { {ImageNet: A Large-Scale Hierarchical Image Database} },
        BOOKTITLE = {CVPR09},
        YEAR = {2009},
        BIBSOURCE = ""}