DeepFruit Dataset
Dataset of varied, annotated and structured fruit images for recognition, classification, and applications in dietary management.
21,122 JPEG images, divided into sets train (16,899) and test (4,223), annotations in CSV form
CC BY 4.0
Description
The DeepFruit Dataset includes 21,122 JPEG images of 20 types of fruit, accompanied by accurate annotations in CSV format. The images are divided into training and test sets to facilitate model validation.
What is this dataset for?
- Train fruit classification and image recognition models
- Develop systems for estimating nutrition, calories, and composition
- Applications in dietetics, food planning and public health
Can it be enriched or improved?
Yes, it is possible to add detailed annotations on nutritional characteristics, to integrate data on maturity or quality, and to extend the corpus with more fruit varieties.
🔎 In summary
🧠 Recommended for
- Vision researchers
- ML developers for dietetics
- Public health projects
🔧 Compatible tools
- PyTorch
- TensorFlow
- OpenCV
- Pandas
- Scikit-learn
💡 Tip
Use CSV files for effective label management during preprocessing.
Frequently Asked Questions
What are the main fruit categories in this dataset?
The dataset contains 20 different types of fruit, each clearly annotated in the CSV files.
Is this dataset ready to use to train a classification model?
Yes, it is structured in train/test sets and provides ready-to-use annotations.
What format do images and annotations take?
The images are in JPEG, the annotations in CSV format associating each image with its fruit label.
