Araştırma Makalesi
BibTex RIS Kaynak Göster

Cycle-GAN ile Modern İç Mekânların Bilim Kurgu Ortamları Olarak Yeniden Üretilmesi

Yıl 2020, Cilt: 1 Sayı: 3, 71 - 94, 30.09.2020

Öz

Derin öğrenme, karmaşık görevlerin ve büyük veri setlerinin işlenmesi gereken problemlerde, yapay sinir ağlarını kullanan bir makine öğrenmesi yöntemidir. Derin öğrenme ile daha önceden uzman bir insan tarafından bilgisayara aktarılması gereken veriye ait özelliklerin, salt bilgisayar tarafından işlenmesi mümkün hale gelmiştir. Derin öğrenmenin alt sistemlerinden biri olan Üretken Rakip Yapay Sinir Ağları (GAN) algoritması, birbirine zıt çalışan iki sinir ağının birbiri ile çekişmesinden faydalanmaktadır. Üretici ağ gerçek olmayan görseller üretirken, ayırt edici ağ, üretilen görselleri değerlendirmekte ve görselin sahte veya gerçek olduğu bilgisini üretmektedir. İki ağ arasındaki bu çekişmeli durum ayırt edici ağın gerçek ile sahteyi ayıramayacağı kadar kaliteli görseller üretilene kadar tekrarlanmaktadır. Bu nedenle GAN algoritması özellikle görüntü işleme ve görüntü çeviri problemlerinde tercih edilmektedir. Derin öğrenmenin sunduğu görüntü işleme teknikleri ile karmaşık mekânsal verilerin kurgulanması ve mekânsal kurguların görüntüler üzerinden tekrar üretimi mümkündür. Bu çalışmanın amacı, farklı özelliklere sahip iç mekânların bir bilim kurgu filminin parçası olma durumunu ve bu durumdan türeyen yeni mekânsal potansiyelleri araştırmaktır. Bu bağlamda görüntü işleme için uygun bir teknik olan GAN algoritması kullanılarak, modern iç mekânlar bilim kurgu mekânları olarak yeniden yorumlanmıştır. Çalışmada modern iç mekân fotoğraflarından ve bilim kurgu filmlerinden olmak üzere iki farklı veri seti oluşturulmuştur. Böylece modern iç mekânların bilim kurgu filmlerinde yer alması durumunda söz konusu mekânların morfolojik olarak nasıl yorumlanabileceği araştırılmıştır.

Kaynakça

  • Abdulkader A., Lakshmiratan A., and Zhang J. (2016) Introducing DeepText: Facebook’s text understanding engine. Erişim Adresi: https://tinyurl.com/jj359dv
  • As, I., Pal, S., & Basu, P. (2018). Artificial intelligence in architecture: Generating conceptual design via deep learning. International Journal of Architectural Computing, 16(4), 306–327.
  • Borysiuk, Zbigniew & Konieczny, Mariusz & Kręcisz, Krzysztof & Pakosz, Paweł. (2018). Application of sEMG and Posturography as Tools in the Analysis of Biosignals of Aging Process of Subjects in the Post-production Age. 10.1007/978-3-319-75025-5_3.
  • De Haan H. (1988). Architects in competition: international architectural competitions of the last 200 years. London:Thames and Hudson.
  • Duda, Richard O.; Hart, Peter E.; Stork, David G. (2001). "Unsupervised Learning and Clustering". Pattern classification (2nd ed.). Wiley. ISBN 0-471-05669-3.
  • Choi, Y., Choi, M., Kim, M., Ha, J. W., Kim, S., & Choo, J. (2018). Stargan: Unified generative adversarial networks for multi-domain image-to-image translation. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 8789-8797).
  • Gero, J. S. (1996). Artificial intelligence in computer-aided design: Progress and prognosis.
  • Goodfellow, I., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., & Bengio, Y. (2014). Generative adversarial nets. In Advances in neural information processing systems (pp. 2672-2680).
  • Hitaj, B., Ateniese, G., & Perez-Cruz, F. (2017, October). Deep models under the GAN: information leakage from collaborative deep learning. In Proceedings of the 2017 ACM SIGSAC Conference on Computer and Communications Security (pp. 603-618).
  • Hopfield, J. J. (1982). "Neural networks and physical systems with emergent collective computational abilities". Proc. Natl. Acad. Sci. U.S.A. 79 (8): 2554–2558. Bibcode:1982PNAS...79.2554H. doi:10.1073/pnas.79.8.2554. PMC 346238. PMID 6953413
  • Huang, X., Liu, M. Y., Belongie, S., & Kautz, J. (2018). Multimodal unsupervised image-to-image translation. In Proceedings of the European Conference on Computer Vision (ECCV) (pp. 172-189).
  • Isola, P., Zhu, J. Y., Zhou, T., & Efros, A. A. (2017). Image-to-image translation with conditional adversarial networks. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 1125-1134).
  • Mitchell, T. M. (1997). Machine learning. Singapore: McGraw-Hill.
  • Nilsson, N. T. (1980). Machine Learning Principles of Artificial Intelligence.
  • Simon, Herbert A. (1973). The Structure of Ill-Defined Problems. Artificial Intelligence. Sayı 4, Bölüm 3-4, sayfa: 181-201.
  • Tamke, M., Nicholas, P., & Zwierzycki, M. (2018). Machine learning for architectural design: Practices and infrastructure. International Journal of Architectural Computing, 16(2), 123-143.
  • Zhu, J. Y., Park, T., Isola, P., & Efros, A. A. (2017). Unpaired image-to-image translation using cycle-consistent adversarial networks. In Proceedings of the IEEE international conference on computer vision (pp. 2223-2232).
  • URL-1: https://www.freecodecamp.org/news/an-intuitive-introduction-to-generative-adversarial-networks-gans-7a2264a81394/
  • URL-2: https://towardsdatascience.com/ai-architecture-f9d78c6958e0
  • URL-3: https://www.architecture.com/image-library/ribapix.html?keywords=modern%20interiors.

Regenerating Modern Interiors into Science Fiction Environments via Cycle-GAN

Yıl 2020, Cilt: 1 Sayı: 3, 71 - 94, 30.09.2020

Öz

Deep learning is a machine learning method that uses artificial neural networks for complex tasks and problems that require the processing of large data sets. Deep learning has shown that it is possible to process the properties of the data that previously needed to be transferred to the computer by an expert person, only by a computer. Generative Adversarial Networks (GAN) algorithm, one of the subsystems of deep learning, takes advantage of the contention of two neural networks working opposite each other. While the Generator produces fake images, the Discriminator evaluates the images and generates the information that the image is fake or real. This contentious situation between two networks repeats until the Discriminator cannot distinguish the image is fake. For this reason, researchers prefer to use the GAN especially in image processing and image translation problems. With the image processing techniques offered by deep learning, it is possible to process complex spatial data and to reproduce spatial fictions through images. The study aims to investigate the new spatial potentials of interior spaces with different characteristics. In this context, modern interiors are reinterpreted as science fiction spaces by using the GAN algorithm, which is a suitable technique for image processing. In this study, we created two different data sets from modern interior photographs and science fiction movies. Thus, we tried to investigate how modern interiors can change morphologically when they become a part of science fiction movies.

Kaynakça

  • Abdulkader A., Lakshmiratan A., and Zhang J. (2016) Introducing DeepText: Facebook’s text understanding engine. Erişim Adresi: https://tinyurl.com/jj359dv
  • As, I., Pal, S., & Basu, P. (2018). Artificial intelligence in architecture: Generating conceptual design via deep learning. International Journal of Architectural Computing, 16(4), 306–327.
  • Borysiuk, Zbigniew & Konieczny, Mariusz & Kręcisz, Krzysztof & Pakosz, Paweł. (2018). Application of sEMG and Posturography as Tools in the Analysis of Biosignals of Aging Process of Subjects in the Post-production Age. 10.1007/978-3-319-75025-5_3.
  • De Haan H. (1988). Architects in competition: international architectural competitions of the last 200 years. London:Thames and Hudson.
  • Duda, Richard O.; Hart, Peter E.; Stork, David G. (2001). "Unsupervised Learning and Clustering". Pattern classification (2nd ed.). Wiley. ISBN 0-471-05669-3.
  • Choi, Y., Choi, M., Kim, M., Ha, J. W., Kim, S., & Choo, J. (2018). Stargan: Unified generative adversarial networks for multi-domain image-to-image translation. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 8789-8797).
  • Gero, J. S. (1996). Artificial intelligence in computer-aided design: Progress and prognosis.
  • Goodfellow, I., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., & Bengio, Y. (2014). Generative adversarial nets. In Advances in neural information processing systems (pp. 2672-2680).
  • Hitaj, B., Ateniese, G., & Perez-Cruz, F. (2017, October). Deep models under the GAN: information leakage from collaborative deep learning. In Proceedings of the 2017 ACM SIGSAC Conference on Computer and Communications Security (pp. 603-618).
  • Hopfield, J. J. (1982). "Neural networks and physical systems with emergent collective computational abilities". Proc. Natl. Acad. Sci. U.S.A. 79 (8): 2554–2558. Bibcode:1982PNAS...79.2554H. doi:10.1073/pnas.79.8.2554. PMC 346238. PMID 6953413
  • Huang, X., Liu, M. Y., Belongie, S., & Kautz, J. (2018). Multimodal unsupervised image-to-image translation. In Proceedings of the European Conference on Computer Vision (ECCV) (pp. 172-189).
  • Isola, P., Zhu, J. Y., Zhou, T., & Efros, A. A. (2017). Image-to-image translation with conditional adversarial networks. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 1125-1134).
  • Mitchell, T. M. (1997). Machine learning. Singapore: McGraw-Hill.
  • Nilsson, N. T. (1980). Machine Learning Principles of Artificial Intelligence.
  • Simon, Herbert A. (1973). The Structure of Ill-Defined Problems. Artificial Intelligence. Sayı 4, Bölüm 3-4, sayfa: 181-201.
  • Tamke, M., Nicholas, P., & Zwierzycki, M. (2018). Machine learning for architectural design: Practices and infrastructure. International Journal of Architectural Computing, 16(2), 123-143.
  • Zhu, J. Y., Park, T., Isola, P., & Efros, A. A. (2017). Unpaired image-to-image translation using cycle-consistent adversarial networks. In Proceedings of the IEEE international conference on computer vision (pp. 2223-2232).
  • URL-1: https://www.freecodecamp.org/news/an-intuitive-introduction-to-generative-adversarial-networks-gans-7a2264a81394/
  • URL-2: https://towardsdatascience.com/ai-architecture-f9d78c6958e0
  • URL-3: https://www.architecture.com/image-library/ribapix.html?keywords=modern%20interiors.
Toplam 20 adet kaynakça vardır.

Ayrıntılar

Birincil Dil Türkçe
Konular Yazılım Testi, Doğrulama ve Validasyon, Mimarlık
Bölüm Araştırma Makaleleri
Yazarlar

Esra Yağdır Çeliker Bu kişi benim 0000-0002-1817-3829

Gizem Efendioğlu

Özgün Balaban Bu kişi benim 0000-0002-7270-2058

Yayımlanma Tarihi 30 Eylül 2020
Yayımlandığı Sayı Yıl 2020 Cilt: 1 Sayı: 3

Kaynak Göster

APA Çeliker, E. Y., Efendioğlu, G., & Balaban, Ö. (2020). Cycle-GAN ile Modern İç Mekânların Bilim Kurgu Ortamları Olarak Yeniden Üretilmesi. Journal of Computational Design, 1(3), 71-94.

88x31.png

JCoDe makaleleri "Creative Commons Attribution-NonCommercial 4.0 International License" altında yayınlanmaktadır.