This project is an AI-based Android application to recognize the fruit via a phone camera.
Implementation has training Convolutional Neural Network (CNN) and designing interface steps.
A CNN first is trained with fruit images.
During training, different augmentation techniques are utilized to get a better model.
The trained then model is saved as a TensorFlow Lite model to use in the application.
I designed the interface of the application in Android Studio.
The application shows the camera scene and the three highest prediction results.
Finally, the application with the trained model is set up on the Android device.
The application can recognize the fruit via a phone camera in real-time.
