How to determine if tomatoes are ripe
A neural network that determines if tomatoes are ripe or not. It includes both detection and classification.
Client
Vladislav Z.
Service
Ripe Tomatoes NN Detector
Date
May 5, 2021
Challenge
Make a neural network that detects tomatoes in the image and classifies them into ripe and not ripe.
Solution
The architecture of the neural network is chosen, the dataset is assembled, augmented and annotated. The neural network is trained and its metrics analysed. Method for model inference is prepared.
The dataset was compiled using images from the Internet. In total, 500 images of tomatoes were collected. To increase the number of images, augmentation was applied. So, I enlarged the number of pictures in 1.5 times.
The roboflow service was used to annotate the dataset:
For the inference, a method was written in Python and a GUI was prepared.
Below you can see examples of how the program works:
