Classification of Benjapakee Buddha amulets image by deep learning
Article Sidebar
Main Article Content
Abstract
Buddha amulet is a symbol of belief and religion, which has a magical power and rare sacred object. Some Buddha amulets can only be seen by those who have not studied cannot know the name or identify the Buddha amulet. Nowadays, the image recognition using machine learning technology has developed rapidly. Deep learning is one of the technology that helps computers to recognize images precisely. The development of appropriate model for the image recognition might lead to create an automatic system for classification of Benjapakee Buddha amulets. The research presented the applied Convolutional Neural Network system (CNN) model of deep learning to classify the Benjapakee Buddha Amulets images. The aim of the research was to develop this suitable architecture deep learning for recognizing the Benjapakee Buddha amulets accurately for the people who did not know and identify these amulets. The experiment consisted of the collection of 100 amulet images for each type, a total of 500 images, and then the test was conducted with a fitty new sample. The results showed that the 3-layer convolution architecture produced the best results. The efficiency of this model could correctly identify 80% of Phra Somdej, Phra Nang Phaya, Phra Rod and Phra Somgor, and 70% of Phra Phong images. The differences between the images of the Benjapakee Buddha amulet of the Buddha image have distinct characteristics, making the ability to accurately identify the images. However, the accuracy of the test depended on the number of tested sample images. The proposed model may only be suitable for images tested, so those interested in image recognition may have to further develop a model suitable for the images.
Article Details
Published manuscript are the rights of their original owners and RMUTSB Academic Journal. The manuscript content belongs to the authors' idea, it is not the opinion of the journal's committee and not the responsibility of Rajamangala University of Technology Suvarnabhumi
References
Fugthong, P., & Meesad, P. (2013). Buddha amulet information retrieval using digital images combined with feature extraction and K-nearest neighbor techniques. Information Technology Journal, 9(2), 34-40. (in Thai)
Korbuakaew, C. (2007). Identification of amulets with special feature matching (Master’s thesis). Silpakorn University, Bangkok. (in Thai)
Krizhevsky, A., Sutskever, I., & Hinton, G. (2012). ImageNet classification with deep convolutional neural networks. In F. Pereira (Ed.), NIPS’12 Proceedings of the 25th International Conference on Neural Information Processing Systems (pp. 1097-1105). Nevada: Curran Associates Inc.
LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. Nature International journal of science, 521, 436-444.
Lin, M., Chen, Q., & Yan, S. (2014). Network in network. In Y. Bengio (Ed.), International Conference on Learning Representations 2014 (ICLR 2014). Banff: ICLR.
Srivastava, N., Hinton, G., Krizhevsky, A., Sutskever, I., & Salakhutdinov, R. (2014). Dropout: a simple way to prevent neural networks from overfitting. J. Mach. Learn. Res., 15, 1929-1958.
Noord, N., & Postma, E. (2016). Learning scale-variant and scale-invariant features for deep image classification. Retrieved 7 April 2019, from https:// arXiv:1602.01255.pdf
Orlando, J. I., Prokofyeva, E., del Fresno, M., & Blaschko, M. B. (2017). An ensemble deep learning based approach for red lesion detection in fundus images. Comput Methods Programs Biomed, 153, 115-127.
Phawattanakul, K. (2012). Suggestion Mining from reviewers’ reviews of television programs (Master’s thesis). National Institute of Development Administration.
Phinitsap, S. (2012). Education for the sake of creativity: Case the degeneration in the faith on the Buddha series five parties in Thai society (Master’s thesis). Srinakharinwirot University, Bangkok.
Remez, T., Litany, O., Giryes, R., & Bronstein, A. M. (2017). Deep convolution denoising of low-light images. Retrieved 7 April 2019, from https://arxiv.org/pdf/1701.01687.pdf
Sivaramakrishnan, R., Antani, S., & Jaeger, S. (2017). Visualizing Deep Learning Activations for Improved Malaria Cell Classification. In Proceedings of First workshop Medical Informatics and Healthcare, Machine Learning Research vol.69 (pp. 40-47). Halifax Canada: Machine Learning Research.
Springenberg, J. T., Dosovitskiy, A., Brox, T. & Riedmiller, M. (2015). Striving for simplicity: The all convolutional net. In Y. Bengio (Ed.), International Conference on Learning Representations 2015 (ICLR 2015) Workshop Track. San Diego: ICLR.
Thai Buddha Image Admiration Association. (2019). Genesis of the Benjapakee Buddha Amulet. Retrieved 7 April 2019, from https://www.samakomphra.com (in Thai)