Fruits Detection using CNN model.
Dataset used :
Fruits 360
A dataset of images consists of various fruits and vegetables.
About Dataset
- Total number of images: 90483.
- Training set size: 67692 images (one fruit or vegetable per image).
- Test set size: 22688 images (one fruit or vegetable per image).
- Multi-fruits set size: 103 images (more than one fruit (or fruit class) per image).
- Number of classes: 131 (fruits and vegetables).
- Image size: 100x100 pixels.
You can have this dataset from Kaggle.