|
Improving rotten fruit classification accuracy through fusion of multiple pretrained CNN models |
|---|---|
| รหัสดีโอไอ | |
| Creator | 1. Singgih Tulus Makmud 2. Adrian Natanael Haryanto 3. Simeon Yuda Prasetyo |
| Title | Improving rotten fruit classification accuracy through fusion of multiple pretrained CNN models |
| Publisher | Faculty of Engineering, Khon Kaen University |
| Publication Year | 2567 |
| Journal Title | Engineering and Applied Science Research |
| Journal Vol. | 51 |
| Journal No. | 3 |
| Page no. | 313-320 |
| Keyword | Rotten fruit classification, CNN, Fusion model, Transfer learning, Fine-tuning |
| URL Website | https://ph01.tci-thaijo.org/index.php/easr/index |
| Website title | Engineering and Applied Science Research |
| ISSN | 2539-6161 |
| Abstract | Agriculture has been the main source of livelihood in Indonesia. The main sector lives in fruit sales, which are important for our healthcare. That is why it is important to keep the fruits fresh into the market. To address this issue, machine learning can help identify rotten fruits with deep learning techniques to differentiate fresh fruit from rotten fruit. Some deep learning techniques will be used in this classification, such as VGG16, MobileNetV2, Xception, ResNet-50, and InceptionV3, which are the common models used in fruit classification in some research. Every model will use fine tuning to increase its performance. The fusion model method will also be used in the classification to improve the performance of the model. Lastly, all of the models will be compared to see which is the best model for classification. In the end, there is a model with the best accuracy, which is MobileNetV2, with 0.99706 accuracy before the fusion model. The result of the fusion model gives the best accuracy, with 0.99926 accuracy. These results showed that the fusion model was best used for fruit classification and can help supermarkets classify rotten and fresh fruits, preventing more issues in the future. Still, this paper did not represent all of the machine learning classification algorithms. Further research is needed to compare the performance of different deep learning approaches in detecting rotten fruit in supermarkets. |
