Selection Of The Most Suitable Parameters For Classification Of Hazelnut Fruit Using Shufflenet Deep Learning Algorithm
Published in BZT Turan, 2025
This study investigates the use of the ShuffleNet deep learning algorithm to classify hazelnut fruits into good, bad, and inner categories, achieving high accuracy with a dataset of 15,770 images. Among the optimizers tested, ShuffleNet with ADAM demonstrated the best performance, with 99.89% training accuracy, 99.96% test accuracy, and superior metrics including 0.9997 precision, 1 specificity, and 0.9998 F1-score.