| Yazarlar |
Doç. Dr. Kemal AKYOL
Kastamonu Üniversitesi, Türkiye |
Abdulkadir Karacı
Samsun University, Turkey |
| Özet |
| The successful management of solid waste, which has become a major issue in urban life, reduces environmental pollution. To address this issue, numerous studies using various deep learning models have been conducted. This study focuses on traditional convolutional neural networks and transformer-based deep learning architectures and provides a detailed examination summary. In this context, many experimental studies dealing with the PoolFormer and ResNet transformers and the DenseNet201 and ResNet50 pre-trained CNN models were run on the TrashNet dataset containing waste images. The Random Forest (RF) classifier was trained on different deep features extracted from 80% of this dataset, and the remaining 20% was reserved as the test dataset. Test dataset was used to validate the performance of the RF. According to the results, RF with DenseNet201 deep features presents 96.4% classification accuracy. Furthermore, the RF-based hybrid model fed with the trained DenseNet201 outperformed other studies using the same dataset in the literature in terms of classification accuracy. The hybrid method employed used in this study may serve as a model for future research. |
| Anahtar Kelimeler |
| DenseNet201 deep features | Random Forest | Recycling | Waste classification |
| Kitap Adı | Lecture Notes on Data Engineering and Communications Technologies |
| Bölüm(ler) | Waste Classification Using Random Forest Classifier with DenseNet201 Deep Features |
| Kitap Türü | Kitap Bölümü |
| Kitap Alt Türü | Alanında uluslararası yayımlanan kitap bölümü |
| Kitap Niteliği | Scopus indeksinde taranan bilimsel kitap |
| Kitap Dili | İngilizce |
| Basım Tarihi | 01-2023 |
| ISBN | 0 |
| Basıldığı Ülke | |
| Basıldığı Şehir |
| Atıf Sayıları |
