Performance Evaluation of the Decision Level Fusion in Dried-Nut Species Classification

Abstract

The proposed study investigates the use of ResNet50 and DenseNet201 networks, which are deep learning network architectures, to feature extraction from the dataset consisting of 11-class dried-nuts images within the scope of transfer learning and to classify products with high accuracy with support vector machines over the obtained feature sets. In addition, the effect of the new feature dataset created by combining the features obtained from two different pre-trained networks with the feature-level fusion approach on the classification performance was also examined within the scope of the study. For the validation of the results, the experiments were carried out under the 5-fold cross-validation technique. When the classification results are examined, classification accuracies of 97.86%, 98.09% and 98,68% were obtained, respectively, as a result of the classification of the extracted features using the ResNet50, DenseNet201 and Fusion architectures with linear core support vector machines. When the feature-level fusion approach was applied, it was observed that the classification accuracy increased to 98.68%.

Keywords

• Dried-Nuts
• Classification
• Feature Extraction.
• DenseNet201
• ResNet50

Öznitelik Seviyesinde Füzyon Yaklaşımının Kuruyemiş Tür Sınıflandırılmasında Performans Değerlendirmesi

Abstract

Önerilen çalışma, derin öğrenme ağ mimarilerinden ResNet50 ve DenseNet201 ağlarının öğrenme aktarımı kapsamında 11 sınıflı kuruyemiş görüntülerinden oluşan veri setinden anlamlı özelliklerin çıkarılmasında kullanılmasını ve elde edilen özellik kümeleri üzerinden karar destek makineleri ile ürünlerin yüksek doğrulukta sınıflandırılmasını araştırmaktadır. Ayrıca çalışma kapsamında özellik seviyesi füzyonu yaklaşımıyla, iki farklı ön eğitimli ağdan elde edilen özelliklerin birleştirilmesi ile oluşturulan yeni özellik veri kümesinin, sınıflandırılma performansına olan etkisi de incelenmiştir. Sonuçların validasyonu için deneyler 5 katlı çapraz doğrulama tekniği kapsamında gerçekleştirilmiştir. Sınıflandırma sonuçları incelendiğinde, ResNet50 ve DenseNet201, Füzyon mimarileri kullanılarak çıkarılan özelliklerin doğrusal çekirdekli karar destek makineleri ile sınıflandırılması neticesinde sırasıyla %97,86, %98,09 ve %98,68 sınıflandırma doğrulukları elde edilmiştir.

Keywords

• Sınıflandırma
• DenseNet201
• ResNet50
• Özellik Çıkarımı.
• Kuruyemiş

References

1. Ecemiş, İ. N., & İlhan, H. O. (2023). The performance comparison of pre-trained networks with the proposed lightweight convolutional neural network for disease detection in tomato leaves. Journal of the Faculty of Engineering and Architecture of Gazi University, 38(2), 693-705.
2. Vidyarthi, S. K., Singh, S. K., Tiwari, R., Xiao, H. W., & Rai, R. (2020). Classification of first quality fancy cashew kernels using four deep convolutional neural network models. Journal of Food Process Engineering, 43(12), e13552.
3. Dheir, I. M., Mettleq, A. S. A., Elsharif, A. A., & Abu-Naser, S. S. (2020). Classifying nuts types using convolutional neural network. International Journal of Academic Information Systems Research (IJAISR), 3(12).
4. Costa, L., Ampatzidis, Y., Rohla, C., Maness, N., Cheary, B., & Zhang, L. (2021). Measuring pecan nut growth utilizing machine vision and deep learning for the better understanding of the fruit growth curve. Computers and Electronics in Agriculture, 181, 105964.
5. Wang, B., Li, H., You, J., Chen, X., Yuan, X., & Feng, X. (2022). Fusing deep learning features of triplet leaf image patterns to boost soybean cultivar identification. Computers and Electronics in Agriculture, 197, 106914.
6. Jan, R., Kour, H., Manhas, J., & Sharma, V. Recognition of Dry Fruits using Deep Convolutional Neural Network.
7. Villacrés, J. F., & Auat Cheein, F. (2020). Detection and characterization of cherries: A deep learning usability case study in Chile. Agronomy, 10(6), 835.
8. Mao, S., Li, Y., Ma, Y., Zhang, B., Zhou, J., & Wang, K. (2020). Automatic cucumber recognition algorithm for harvesting robots in the natural environment using deep learning and multi-feature fusion. Computers and Electronics in Agriculture, 170, 105254.

9. Han, Y., Liu, Z., Khoshelham, K., & Bai, S. H. (2021). Quality estimation of nuts using deep learning classification of hyperspectral imagery. Computers and Electronics in Agriculture, 180, 105868.
10. Wang, Z., Jin, L., Wang, S., & Xu, H. (2022). Apple stem/calyx real-time recognition using YOLO-v5 algorithm for fruit automatic loading system. Postharvest Biology and Technology, 185, 111808.
11. Varshni, D., Thakral, K., Agarwal, L., Nijhawan, R., & Mittal, A. (2019, February). Pneumonia detection using CNN based feature extraction. In 2019 IEEE international conference on electrical, computer and communication technologies (ICECCT) (pp. 1-7). IEEE.
12. An, R., Perez-Cruet, J., & Wang, J. (2022). We got nuts! use deep neural networks to classify images of common edible nuts. Nutrition and Health, 02601060221113928.
13. Özkaya, U., Öztürk, Ş., & Barstugan, M. (2020). Coronavirus (COVID-19) classification using deep features fusion and ranking technique. In Big Data Analytics and Artificial Intelligence Against COVID-19: Innovation Vision and Approach (pp. 281-295). Springer, Cham.