Determining Banana Ripeness Using MobileNet

Bananas are one of the most popular and economically important fruit crops in the world. Not only do they have a sweet taste, but they are also rich in various nutrients that are useful for human body. To find an efficient method for classifying banana ripeness, images of bananas were used to train...

Full description

Saved in:
Bibliographic Details
Published in2024 12th International Electrical Engineering Congress (iEECON) pp. 1 - 6
Main Authors Chen, Hong, Phoophuangpairoj, Rong
Format Conference Proceeding
LanguageEnglish
Published IEEE 06.03.2024
Subjects
banana ripeness
Biological system modeling
CNN
Crops
Electrical engineering
image processing
Image recognition
Machine learning
MobileNet
Neural networks
ResNet50
Training
Online AccessGet full text

Cover

More Information
Summary:Bananas are one of the most popular and economically important fruit crops in the world. Not only do they have a sweet taste, but they are also rich in various nutrients that are useful for human body. To find an efficient method for classifying banana ripeness, images of bananas were used to train the MobileNet, ResNet50, and Convolution Neural Network (CNN) machine learning classifiers. These machine learning classifiers can automatically extract the features from the trained images and then use them to build models for classification. The results indicated that the MobileNet, ResNet50 and CNN classifiers could obtain the accuracies of 98.43%, 94.19% and 95.35%, respectively. The accuracies of MobileNet classification from class 0 to class 6 were 99.62%, 98.85%, 99.23%, 100%, 100%, 99.62% and 99.62%, respectively.
DOI:10.1109/iEECON60677.2024.10537906