• Fruits-360 original-size (Original size images. Currently contains 24 fruits and vegetables)

• Fruits-360 100x100 pixels (Images scaled to 100x100 pixels. Currently contains 90380 pictures of 131 fruits, vegetables and nuts)

• Fruits-360 metadata (Attributes. Currently contains data of 26 fruits and vegetables)

Mihai Oltean, Fruits-360 dataset, 2017-.

• 2016 (begining of). I met with a private company that needed a robot to handle the goods from their food stores. My solution was to use Jenny 5 robot for this purpose. Even if Jenny 5 robot has three Logitech C910 cameras, we had no serious algorithm capable of recognizing foods. So, the first step was to build a dataset with many images of foods that could be fed into a neural network as training data ...

• 2016 (at the end). I started collecting fruits (either from my own garden or purchased from local stores). I stuck them into a low-rpm motor and filmed them. Initially, the films were made inside my room, which has blue-painted walls, whose reflection is visible in many images from that time.

• 2017 (begining of). A student of mine wrote an example of using TensorFlow for the Fruits-360 dataset.

• 2017 (summer of). The student presented his work at a local symposium for students.

• 2017 (end of). A more complete paper with the dataset and numerical experiments has been accepted and publised by Acta Univ. Sapientiae. The paper has been updated several times after publication to include results for newer dataset versions.

• 2017 (end of). The dataset has been uploaded to Kaggle: https://www.kaggle.com/datasets/moltean/fruits.

• 2021 (begining of). Original (captured) size images will be published from now on.

• 2021 (begining of). The metadata (properties) of fruits will be included in the dataset.

• 2024.02.17. The dataset has been moved to a new location on GitHub: https://gihub.com/fruits-360. The organization contains multiple repositories.

• Mihai Oltean, Fruits-360 dataset: new research directions, Technical report, 2021.

• Horea Muresan, Mihai Oltean, Fruit recognition from images using deep learning, Acta Univ. Sapientiae, Informatica Vol. 10, Issue 1, pp. 26-42, 2018.

More papers using Fruits-360 dataset are here ...

MIT license.

You may do whatever you want with this dataset as long as you mention the author.

Mihai Oltean

email: mihai.oltean@gmail.com

All fruits, vegetables, etc, are from:

• my own garden or
• purchased, by me, from local stores.

I have recorded all the movies containing the fruits, vegetables, etc in the Fruits-360 dataset.

I extracted (from movies) all pictures in the Fruits-360 dataset.

I have NOT taken pictures from other sources.

I have NOT generated artificial pictures of fruits, vegetables etc.