Fruit 360 Subset: 10 Fruit and Vegetable Images for Classification

Fruit 360 is a dataset of 90,380 images of 131 fruits and vegetables. This subset contains images for 10 specific fruits and vegetables, with each image being a 100x100 pixel RGB color image. The dataset was created by Kaivan Kamali at Pennsylvania State University using utilities from the original work by H. Muresan and M. Oltean.

Use Cases

• Train image classification models based on 100x100 pixel RGB images of fruits and vegetables.
• Benchmark model performance on a reduced 10-class subset derived from a larger dataset.
• Experiment with feature vector representations (30,000 values per image) for computer vision tasks.

Strengths

• Contains 90,380 total images in the original full dataset.
• Images have a consistent resolution of 100x100 pixels in RGB format.
• Labels for the 10 selected classes are mapped to a numeric range (0-9) for model training.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Last update date is unknown; freshness unverified.
• The geographic and temporal context of the image collection is not specified.

Provenance

Source

Derived from the Fruit 360 dataset on Kaggle, originally from 'Fruit recognition from images using deep learning' by H. Muresan and M. Oltean.

Collection Method

A subset of 10 fruit/vegetable classes was selected from the original dataset, and feature vectors (30,000 values per image) were created.