Classification and Evaluation of Fruit Using a Convolutional Neural Network (CNN)
Team Good Apple:
Project Manager: Michael Vincent Grillo
Project Associates: Matthew Davis, William Ah Tou, Kyle Cloud, Davis Hill, Hanan Kwok
Problem: During multiple stages in produce sales, retailers often receive unlabeled boxes containing fruit, some of which does not meet quality standards set forth by the company. Sorting fruit by hand can be expensive and time-consuming.
Solution: A CNN model trained to distinguish fresh apples, fresh oranges, fresh bananas, rotten apples, rotten oranges or rotten bananas to increase quality control efficiency for the fruit retailer. By leveraging Convolutional Neural Networks, a model was created that accomplished this goal with a high level of accuracy.