cub200_dataset.py

# -*- coding: utf-8 -*-
"""cub200_dataset.py

Automatically generated by Colaboratory.

Original file is located at
    https://colab.research.google.com/drive/1qC5RnFLP3_9X50ripGf5YtfXnugxBj2m
"""

from PIL import Image
import os
import pandas as pd
from datasets import DatasetDict, DatasetInfo, Features, Value, Sequence, Image, SplitGenerator, GeneratorBasedBuilder, Version

_CITATION = """\
@techreport{WahCUB_200_2011,
  Title = {The Caltech-UCSD Birds-200-2011 Dataset},
  Author = {Wah, C. and Branson, S. and Welinder, P. and Perona, P. and Belongie, S.},
  Year = {2011},
  Institution = {California Institute of Technology},
  Number = {CNS-TR-2011-001}
}
"""

_DESCRIPTION = """\
The CUB-200-2011 dataset contains 11,788 photos of 200 bird species. Each photo comes with detailed annotations, including part locations, bounding boxes, and attributes for studying fine-grained visual categorization.
"""

_HOMEPAGE = "http://www.vision.caltech.edu/visipedia/CUB-200-2011.html"

_DATASET_PATH = "/content/drive/My Drive/cub200/CUB_200_2011"

class CUB2002011(datasets.GeneratorBasedBuilder):
    """CUB-200-2011 dataset for bird species image classification."""

    # Version of the dataset
    VERSION = datasets.Version("1.0.0")

    # Define the features of the dataset, including the image and the label
    def _info(self):
        return datasets.DatasetInfo(
            description="CUB-200-2011 is an image dataset with photos of 200 bird species.",
            features=datasets.Features({
                "image": datasets.Image(),
                "label": datasets.ClassLabel(names=[f"species_{i:03d}" for i in range(1, 201)]),
            }),
            supervised_keys=("image", "label"),
            homepage="http://www.vision.caltech.edu/visipedia/CUB-200-2011.html",
            citation="""@techreport{WahCUB_200_2011,
            Title = {The Caltech-UCSD Birds-200-2011 Dataset},
            Author = {Wah, C. and Branson, S. and Welinder, P. and Perona, P. and Belongie, S.},
            Year = {2011},
            Institution = {California Institute of Technology},
            Number = {CNS-TR-2011-001}
          }"""
        )

    # Specify the dataset splits
    def _split_generators(self, dl_manager):
        # Assuming the dataset is pre-downloaded
        dl_manager = DownloadManager.download_and_extract("https://data.caltech.edu/records/65de6-vp158/files/CUB_200_2011.tgz")
        return [
            datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"data_dir": data_dir, "split": "train"}),
            datasets.SplitGenerator(name=datasets.Split.TEST, gen_kwargs={"data_dir": data_dir, "split": "test"}),
        ]

    # Generate examples from the dataset directory
    def _generate_examples(self, data_dir, split):
        # Implement logic to iterate over the dataset and yield examples
        # For simplicity, assuming all images are in the 'images' folder and split is ignored
        species_dirs = [p for p in (data_dir / "images").iterdir() if p.is_dir()]
        for species_dir in species_dirs:
            species_label = species_dir.name
            for image_path in species_dir.glob("*.jpg"):
                # The key can be whatever unique identifier you choose; here we use the image path
                yield image_path.stem, {
                    "image": str(image_path),
                    "label": species_label,
                }