Using Deep Learning Techniques Furniture Image Classification

Öz

Bu makale, mobilya görüntülerinin sınıflandırılması konusunda yapay zeka tekniklerinin kullanılmasını ele almaktadır. Mobilya sektöründe birçok farklı tasarım ve tarz arasından seçim yapmanın zorluğu, tüketiciler ve satıcılar için bir sorun oluşturmaktadır. Makine öğrenimi algoritmaları ve sinir ağları, mobilya görüntülerini otomatik olarak sınıflandırma sürecinde kullanılmaktadır. Makalenin amacı, mobilya görüntülerinin sınıflandırılmasıyla tüketicilerin ve mobilya endüstrisi profesyonellerinin karşılaştığı sorunları çözmektir. Makalede, mobilya görüntülerinin sınıflandırılması için beş farklı evrişimli sinir ağı mimarisi kullanılmıştır: Alexnet, VGGNet-19, DenseNet-201, Squeezenet1.1 ve ResNet-152. Bu mimarilerin kullanımıyla %98.87 sınıflandırma başarısı elde edilmiştir. Beş farklı mobilya kategorisi (yatak, sandalye, kanepe, döner koltuk ve masa) sınıflandırılmış ve ResNet-152 mimarisiyle %99.96 ROC (Receiver Operating Characteristic) değeri elde edilmiştir. Ayrıca, transfer öğrenme yaklaşımının kullanılmasıyla daha hızlı ve doğru sonuçlar elde edildiği belirtilmiştir. VGG-19 ve SqueezeNet1.1 mimarileri %97.07 ortalama sınıflandırma doğruluğu sağlarken, en düşük doğruluğu Alexnet modeli (%94.15) gerçekleştirmiştir. Derin öğrenme algoritmalarının kullanılmasıyla görüntülerin özellikleri çıkarılmakta ve sınıflandırılmaktadır. Bu çalışma, teknolojinin daha akıllı ve kullanıcı odaklı bir alışveriş deneyimi sunma potansiyeline sahip olduğunu göstermektedir. Aynı zamanda, mobilya üretim ve satışında verimliliği artırarak rekabet avantajı sağlayabilecek bir mobilya sınıflandırma yöntemi sunmaktadır. Çalışmada elde edilen sonuçlar, mobilya görüntülerinin analizi ve sınıflandırılmasında CNN mimarilerinin etkili olduğu göstermiştir.

Anahtar Kelimeler

• Image Classification
• deep learning
• computer vision
• furniture image
• transfer learning.

Derin Öğrenme Tekniklerini Kullanarak Mobilya Görüntüsü Sınıflandırması

Öz

This paper discusses the use of artificial intelligence (AI) techniques for the classification of furniture images. In the furniture industry, the difficulty of choosing from many different designs and styles poses a problem for consumers and sellers. Machine learning (ML) algorithms and neural networks are used in the process of automatically classifying furniture images. The aim of the paper is to solve the problems faced by consumers and furniture industry professionals with the classification of furniture images. In the paper, five different convolutional neural network architectures are used for furniture image classification: Alexnet, VGGNet-19, DenseNet-201, Squeezenet1.1 and ResNet-152. Using these architectures, 98.87% classification success is achieved. Five different categories of furniture (bed, chair, sofa, swivel chair and table) are classified and an ROC (Receiver Operating Characteristic) value of 99.96% is obtained with the ResNet-152 architecture. In addition, it is reported that faster and more accurate results are obtained by using a transfer learning approach. The VGGNet-19 and SqueezeNet1.1 architectures provided an average classification accuracy of 97.07%, while the Alexnet model (94.15%) achieved the lowest accuracy. Using deep learning algorithms, the features of images are extracted and classified. This study shows that the technology has the potential to deliver a smarter and user-centered shopping experience. It also provides a furniture classification method that can provide a competitive advantage by increasing efficiency in furniture production and sales. The results of the study show that CNN architectures are effective in analyzing and classifying furniture images.

Anahtar Kelimeler

• Görüntü sınıflandırma
• derin öğrenme
• bilgisayar görüşü
• mobilya görüntüsü
• transfer öğrenme

Kaynakça

1. [1] Ting-Ting, S., Ke-Yu, Z., Hui, Z., and Qiao, H., “Interest points guided convolution neural network for furniture styles classification”, In 2019 6th International Conference on Systems and Informatics (ICSAI) (pp. 1302-1307). IEEE, (2019).
2. [2] Ren, S., He, K., Girshick, R., and Sun, J., “Faster r-cnn: Towards real-time object detection with region proposal networks”, Advances in neural information processing systems, 28, (2015).
3. [3] Mo, K., Zhu, S., Chang, A. X., Yi, L., Tripathi, S., Guibas, L. J., and Su, H., “Partnet: A large-scale benchmark for fine-grained and hierarchical part-level 3d object understanding”, In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, (pp. 909-918), (2019).
4. [4] Varvadoukas, T., Giannakidou, E., Gómez, J. V., and Mavridis, N., “Indoor furniture and room recognition for a robot using internet-derived models and object context”, In 2012 10th International Conference on Frontiers of Information Technology, (pp. 122-128). IEEE, (2012).
5. [5] Krizhevsky, A., Sutskever, I., and Hinton, G. E., “Imagenet classification with deep convolutional neural networks”, Communications of the ACM, 60(6), 84-90, (2017).
6. [6] Simonyan, K., and Zisserman, A., “Very deep convolutional networks for large-scale image recognition”, arXiv preprint arXiv:1409.1556, (2014).
7. [7] Xiao, J., Hays, J., Ehinger, K. A., Oliva, A., and Torralba, A., “Sun database: Large-scale scene recognition from abbey to zoo”, In 2010 IEEE computer society conference on computer vision and pattern recognition (pp. 3485-3492). IEEE, (2010).
8. [8] Zhu, B., Yang, C., Yu, C., and An, F., “Product image recognition based on deep learning”, Journal of Computer-Aided Design & Computer Graphics, 30(9), 1778, (2018).

9. [9] Wang, Y., Gao, W., and Wang, Y., “Application of furniture images selection based on neural network”, In AIP Conference Proceedings, (Vol. 1967, No. 1, p. 040016). AIP Publishing LLC, (2018).
10. [10] Zhu, H., Li, X., Huang, W., and Li, C., “Chair Style Recognition and Intelligent Design Method Based on Deep Learning”, furniture, 42, 37-41, (2021).
11. [11] Ye, H., Zhu, X., Liu, C., Yang, L., and Wang, A., “Furniture Image Classification Based on Depthwise Group Over-Parameterized Convolution”, Electronics, 11(23), 3889, (2022).
12. [12] https://www.kaggle.com/datasets/akkithetechie/furniture-detector/download?datasetVersionNumber=1 (accessed on 5 June 2023).
13. [13] Deng, L., and Yu, D., “Deep learning: methods and applications”, Foundations and trends® in signal processing, 7(3–4), 197-387, (2014).
14. [14] Schmidhuber J., “Deep learning in neural networks: An overview”, Neural Networks, 61: 85-117, (2015).
15. [15] Ergün, E., ve Kılıç, K., “Derin öğrenme ile artırılmış görüntü seti üzerinden cilt kanseri tespiti”, Black Sea Journal of Engineering and Science, 4(4), 192-200, (2021).
16. [16] LeCun, Y., Bengio, Y., and Hinton, G., “Deep learning”, Nature, 521(7553), 436-444, (2015).
17. [17] Sharma, N., Jain, V., and Mishra, A., “An analysis of convolutional neural networks for image classification”, Procedia computer science, 132, 377-384, (2018).
18. [18] Krizhevsky, A., Sutskever, I., and Hinton, G. E., “Imagenet classification with deep convolutional neural networks”, Communications of the ACM, 60(6), 84-90., (2017).
19. [19] Simonyan, K., and Zisserman, A., “Very deep convolutional networks for large-scale image recognition”, arXiv preprint arXiv, 1409.1556, (2014).
20. [20] Huang, G., Liu, Z., Van Der Maaten, L., and Weinberger, K. Q., “Densely connected convolutional networks”, In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 4700-4708, (2017).
21. [21] Iandola, F. N., Han, S., Moskewicz, M. W., Ashraf, K., Dally, W. J., and Keutzer, K., “SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and< 0.5 MB model size”, arXiv preprint arXiv, 1602.07360, (2016).
22. [22] He, K., Zhang, X., Ren, S., and Sun, J., “Deep residual learning for image recognition”, In Proceedings of the IEEE conference on computer vision and pattern recognition pp. 770-778, (2016).
23. [23] Pan, S. J., and Yang, Q., “A survey on transfer learning”, IEEE Transactions on knowledge and data engineering, 22(10), 1345-1359, (2010).
24. [24] Du, X., “FISC: Furniture image style classification model based on Gram transformation”, In 2021 3rd International Conference on Advanced Information Science and System, pp. 1-5 (2021).