README.md

metadata

dataset_info:
  features:
    - name: image
      dtype: image
    - name: label
      dtype:
        class_label:
          names:
            '0': no_tomato
            '1': tomato
  splits:
    - name: original
      num_bytes: 8384551
      num_examples: 49
    - name: augmented
      num_bytes: 36005046
      num_examples: 490
  download_size: 44391487
  dataset_size: 44389597
configs:
  - config_name: default
    data_files:
      - split: original
        path: data/original-*
      - split: augmented
        path: data/augmented-*

Dataset Summary

This dataset contains real-world photographs labeled for the presence of tomatoes.
It is designed for binary image classification tasks, where the model predicts whether an image contains a tomato (1) or not (0).

• Original size: 49 images
• Augmented size: 490 images
• Task type: Image Classification (binary)
• Goal: Train models to distinguish between images with and without tomatoes

Data Splits

• No predefined train/test split.
• Users can apply their own strategy (e.g., 80/20 split or k-fold cross-validation).

Intended Uses

• Binary Classification: Distinguish between images containing tomatoes vs. not.
• Computer Vision Training: Baseline dataset for testing CNNs or transfer learning models.
• Educational Use: Demonstrates dataset augmentation in image classification (49 → 490 samples).

Labels

• 0 → Image does not contain tomatoes
• 1 → Image contains tomatoes