FruitNet: Indian Fruits Dataset with Quality

We’ve put together the “FruitNet” dataset to help with recognizing and classifying fruits. It contains over 14,700 top-quality images of six popular Indian fruits. These images are sorted into three folders: “Good quality fruits,” “Bad quality fruits,” and “Mixed quality fruits.”

Each group includes pictures of apples, bananas, guavas, limes, oranges, and pomegranates. The pictures were taken with a mobile phone that has a high-resolution camera, so they vary in backgrounds and lighting.

The “FruitNet” dataset serves as a valuable resource for training, testing, and validating fruit classification or recognition models.

Applications of the FruitNet Dataset

• Automated Sorting Systems: Develops systems for sorting fruits based on quality and type.
• Quality Control: Enhances processes for ensuring fruit quality in production and retail.
• Consumer Information: Provides detailed quality information to consumers for better purchasing decisions.

Challenges in Dataset Development

• Data Collection: Ensures comprehensive and diverse data collection across different regions and seasons.
• Annotation Accuracy: Maintains high standards in annotating quality metrics and defects.
• Integration with Other Data Sources: Combines with other agricultural datasets for holistic analysis.

Future Directions

• Enhanced Data Collection Methods: Utilizes advanced imaging and sensing technologies for better data accuracy.
• Collaboration and Data Sharing: Promotes collaboration among researchers, farmers, and industry stakeholders for continuous improvement.
• Application Expansion: Expands applications to new areas such as precision agriculture and smart farming technologies.