{ "article_id": 1423, "public_id": "dffee4a9-42f3-44d6-ba49-257bd663bbd5", "internal_identifier": null, "source_identifier": "88:8", "status": "published", "type": "research_article", "title": "Üretken Çekişmeli Ağ Kullanarak Eski Görüntüleri Renklendirme ve İyileştirme", "title_secondary": null, "abstract": "Günümüzde derin öğrenme alanı için veri, çalışma sonuçlarının başarı oranı için büyük önem taşımaktadır. Bu ihtiyaca yönelik olarak görüntü işleme konusunda yapabileceğimiz iyileştirmeler ile daha kaliteli ve eksiksiz görüntüler elde etmek ve eski verilerde karşılaştığımız sorunlardan biri olabilen siyah beyaz görüntülerin renkli görüntüye çevrilmesi ve eksik piksellerin tamamlanması bu çalışmadaki amacı oluşturmaktadır. Çalışmada geliştirilen kullanıcı arayüzü sayesinde hem kullanıcıların kişisel görüntüleri düzenlemelerine hem diğer çalışmalarda destekleyici bir araç olarak kullanılmasına hizmet etmektedir. Yöntem olarak ön işleme, derin öğrenme modeli ile eğitim ve rötuş işlemleri sırasıyla kullanılmıştır. Mimari olarak kullandığımız Üretken Çekişmeli Ağ (GAN) çeşitlerinden bir tanesi olan Pix2pix modeli sayesinde kaynak görüntüden hedef görüntüye dönüşüm yapan bir yöntem geliştirilmiştir. Bu bakış açısıyla Pix2pix modeli görüntü iyileştirme ve dönüşümlerimizde ideal bir ağ olmuştur. Deneylerde elde edilen en yüksek başarım sonucu %82 doğruluk oranıyla iç mekân ve dış mekân sınıflandırması için elde edilmiştir. Ayrıca iç mekan renklendirmesi, dış meķan renklendirmesi, insan yüz renklendirmesi, görüntü onarma ve görüntü kalitesini artırma için yapılan deneylerdeki ortalama değerler göz önüne alındığında en yüksek Yapısal Benzerlik Endeks Ölçüsü (SSIM) değeri 0,9256 ve en düşük Tepe Sinyal Gürültü Oranı (PSNR) değeri ise 65,11 dB olmuştur. Çalışmamızda bilimsel bulgulara dayanan tartışma ve değerlendirmelere yer verilmektedir.", "language": "tr", "english_title": "Colorization and Enhancement of Old Images Using Generative Adversarial Networks", "english_abstract": "Nowadays, data for deep learning is crucial to the success rate of the study results. With the improvements we can make to image processing for this need, it is intended to achieve better quality and complete images and to translate black and white images into a color image, which may be one of the problems we face with older data, and to complete missing pixels. Thanks to the user interface developed in the study, it serves both as a user's ability to edit personal images and as a supporting tool for other studies. As a method, pre-processing, deep learning and training and retouching are used in order. One of the varieties of Generative Adversarial Network (GAN) architecture, the Pix2pix model has been developed as a way to transform from source image to target image. From this point of view, the Pix2pix model has been the ideal network for improvement and conversions of images. The highest performance result in experiments was achieved for indoor and outdoor classification with accuracy of 82%. In addition, the highest Structural Similarity Index Measure (SSIM) value was 0.9256 and the lowest Peak Signal to Noise Ratio (PSNR) value was 65.11 dB when the average values were taken into account with the experiments to improve indoor colorization, outdoor colorization, human face colorization, image repair and image quality. Our study includes discussions and assessments based on scientific findings.", "doi": "10.36287/setsci.5.1.008", "doi_url": "https://doi.org/10.36287/setsci.5.1.008", "resource_url": "https://www.set-science.com/a/dffee4a9-42f3-44d6-ba49-257bd663bbd5", "pdf_url": "https://www.set-science.com/a/dffee4a9-42f3-44d6-ba49-257bd663bbd5/pdf", "og_image": "https://www.set-science.com/assets/img/og-card-v2.png", "authors": [ { "public_id": "e25f27c4-6e78-4dbf-abd4-c114286c6e22", "given": "Arda Cem", "family": "Bilecan", "full": "Arda Cem Bilecan", "literal": null, "orcid": null, "orcid_url": null, "email": "ardacembilecan@gmail.com", "corresponding": true, "sequence": "first", "affiliations": [ { "name": "Çanakkale Onsekiz Mart Üniversitesi, Bilgisayar Mühendisliği Bölümü", "department": null, "country": null, "country_code": null, "ror": null, "ror_url": null } ] }, { "public_id": "e9fbb28f-df50-42ed-91bb-9a480f2da3b6", "given": "Simay", "family": "Hoşmeyve", "full": "Simay Hoşmeyve", "literal": null, "orcid": null, "orcid_url": null, "email": "simayhosmeyve@gmail.com", "corresponding": false, "sequence": "additional", "affiliations": [ { "name": "Çanakkale Onsekiz Mart Üniversitesi, Bilgisayar Mühendisliği Bölümü", "department": null, "country": null, "country_code": null, "ror": null, "ror_url": null } ] }, { "public_id": "35da26c1-1bda-430c-b320-b9bddb318cec", "given": "Bahadır", "family": "Karasulu", "full": "Bahadır Karasulu", "literal": null, "orcid": null, "orcid_url": null, "email": "noname@noname.eu", "corresponding": false, "sequence": "additional", "affiliations": [ { "name": "Çanakkale Onsekiz Mart Üniversitesi, Bilgisayar Mühendisliği Bölümü", "department": null, "country": null, "country_code": null, "ror": null, "ror_url": null } ] } ], "keywords": [ "Derin Öğrenme", "Üretken Çekişme Ağlar", "Görüntü İyileştirme", "Sınıflandırma", "Kullanıcı Grafik Arayüzü" ], "translations": [ { "language": "en", "title": "Colorization and Enhancement of Old Images Using Generative Adversarial Networks", "abstract": "Nowadays, data for deep learning is crucial to the success rate of the study results. With the improvements we can make to image processing for this need, it is intended to achieve better quality and complete images and to translate black and white images into a color image, which may be one of the problems we face with older data, and to complete missing pixels. Thanks to the user interface developed in the study, it serves both as a user's ability to edit personal images and as a supporting tool for other studies. As a method, pre-processing, deep learning and training and retouching are used in order. One of the varieties of Generative Adversarial Network (GAN) architecture, the Pix2pix model has been developed as a way to transform from source image to target image. From this point of view, the Pix2pix model has been the ideal network for improvement and conversions of images. The highest performance result in experiments was achieved for indoor and outdoor classification with accuracy of 82%. In addition, the highest Structural Similarity Index Measure (SSIM) value was 0.9256 and the lowest Peak Signal to Noise Ratio (PSNR) value was 65.11 dB when the average values were taken into account with the experiments to improve indoor colorization, outdoor colorization, human face colorization, image repair and image quality. Our study includes discussions and assessments based on scientific findings.", "keywords": [ "Deep Learning", "Generative Adversarial Networks", "Image Enhancement", "Classification", "Graphical User Interface" ] } ], "subjects": [], "subject_names": [], "published_date": "2022-06-01", "published_year": "2022", "received_date": null, "accepted_date": null, "volume": "13", "issue": null, "page_first": "38", "page_last": "43", "article_number": null, "conference": { "short_name": "ISAS2022", "full_title": "5th International Symposium on Innovative Approaches in Smart Technologies", "name": "5th International Symposium on Innovative Approaches in Smart Technologies", "acronym": "ISAS2022", "location": "Online, Türkiye", "start_date": "2022-05-28", "end_date": null, "public_id": "23436f3c-bde1-4f59-b3be-a8ccc770539f" }, "proceedings": { "title": "Proceeding of the 5th International Symposium on Innovative Approaches in Smart Technologies", "series_title": "SETSCI Conference Proceedings", "publisher": "SET Technology", "isbn": null, "eisbn": "978-625-95366-9-9", "issn": null, "eissn": "2687-5527", "doi": null, "doi_prefix": "10.36287", "slug": "setsci-volume-13-2022-legacy-88", "public_id": "04c79b09-4b3a-4deb-84ea-b349d80e3e9e", "url": "https://www.set-science.com/p/setsci-volume-13-2022-legacy-88" }, "license": { "name": "CC BY 4.0", "url": "https://creativecommons.org/licenses/by/4.0/", "code": "CC-BY-4.0", "open_access": true }, "copyright": { "holder": "SET Technology", "year": "2022" }, "funding_statement": null, "funders": [], "references": [ { "order": 1, "raw": "I. J. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu, D. Warde-Farley, S. Ozair, A. Courville, Y. Bengio, “Generative Adversarial Nets”, 2014, arXiv Preprint: arXiv:1406.2661 [stat.ML] https://doi.org/10.48550/arXiv.1406.2661", "doi": null, "title": null, "authors": null, "source": null, "year": null, "volume": null, "pages": null }, { "order": 2, "raw": "M. Mirza, S. Osindero, “Conditional Generative Adversarial Nets”, 2014, arXiv Preprint: arXiv:1411.1784 [cs.LG] https://doi.org/10.48550/arXiv.1411.1784", "doi": null, "title": null, "authors": null, "source": null, "year": null, "volume": null, "pages": null }, { "order": 3, "raw": "E. Learned-Miller, G. B. Huang, A. RoyChowdhury, H. Li, and G. Hua, “Labeled Faces in the Wild: A Survey”. In Advances in Face Detection and Facial Image Analysis, Editörler: M. Kawulok, M. E. Celebi ve B. Smolka, Springer, pp. 189-248, 2016, ISBN: 978-3-319-25958-1", "doi": null, "title": null, "authors": null, "source": null, "year": null, "volume": null, "pages": null }, { "order": 4, "raw": "Labeled Faces in the Wild [Online]: http://viswww.cs.umass.edu/lfw/", "doi": null, "title": null, "authors": null, "source": null, "year": null, "volume": null, "pages": null }, { "order": 5, "raw": "A. Quattoni, A. Torralba, \"Recognizing indoor scenes,\" 2009 IEEE Conference on Computer Vision and Pattern Recognition, 20-25 June 2009, pp. 413-420, doi: 10.1109/CVPR.2009.5206537.", "doi": null, "title": null, "authors": null, "source": null, "year": null, "volume": null, "pages": null }, { "order": 6, "raw": "Kaggle İnternet sitesi [Online]: https://www.kaggle.com", "doi": null, "title": null, "authors": null, "source": null, "year": null, "volume": null, "pages": null }, { "order": 7, "raw": "House Rooms Dataset İnternet sitesi [Online]: https://www.kaggle.com/datasets/robinreni/house-rooms-imagedataset", "doi": null, "title": null, "authors": null, "source": null, "year": null, "volume": null, "pages": null }, { "order": 8, "raw": "Places2 Dataset İnternet sitesi [Online]: http://places2.csail.mit.edu/download.html", "doi": null, "title": null, "authors": null, "source": null, "year": null, "volume": null, "pages": null }, { "order": 9, "raw": "M. Caron, I. Misra, J. Mairal, P. Goyal, P. Bojanowski, A. Joulin, “Unsupervised Learning of Visual Features by Contrasting Cluster Assignments”, 2021: arXiv Preprint arXiv: 2006.09882[cs.CV] https://doi.org/10.48550/arXiv.2006.09882", "doi": null, "title": null, "authors": null, "source": null, "year": null, "volume": null, "pages": null }, { "order": 10, "raw": "O. Ronneberger, P. Fischer, T. Brox, “U-Net: Convolutional Networks for Biomedical Image Segmentation”, 2015, arXiv Preprint: arXiv:1505.04597 [cs.CV],\r\nhttps://doi.org/10.48550/arXiv.1505.04597", "doi": null, "title": null, "authors": null, "source": null, "year": null, "volume": null, "pages": null }, { "order": 11, "raw": "Y. Ho, S. Wookey, “The Real-World-Weight Cross-Entropy Loss Function: Modeling the Costs of Mislabeling”, arXiv Preprint: arXiv:2001.00570 [cs.LG],\r\nhttps://doi.org/10.48550/arXiv.2001.00570", "doi": null, "title": null, "authors": null, "source": null, "year": null, "volume": null, "pages": null }, { "order": 12, "raw": "S. Ruder, “An overview of gradient descent optimization algorithms”, 2017, arXiv Preprint : arXiv:1609.04747 [cs.LG] https://doi.org/10.48550/arXiv.1609.04747", "doi": null, "title": null, "authors": null, "source": null, "year": null, "volume": null, "pages": null }, { "order": 13, "raw": "A. Farhadi, J. Redmon, “YOLOv3: An Incremental Improvement”, 2018, arXiv Preprint: arXiv:1804.02767 [cs.CV], https://doi.org/10.48550/arXiv.1804.02767", "doi": null, "title": null, "authors": null, "source": null, "year": null, "volume": null, "pages": null }, { "order": 14, "raw": "YOLOv3 in PyTorch, Github İnternet Sitesi: https://github.com/ultralytics/yolov3", "doi": null, "title": null, "authors": null, "source": null, "year": null, "volume": null, "pages": null }, { "order": 15, "raw": "G. E. Hinton, A. Krizhevsky, I. Sutskever, “ImageNet Classification with Deep Convolutional Neural Networks”, Advances in Neural Information Processing Systems, Editörler: F. Pereira and C.J. Burges and L. Bottou and K.Q. Weinberger, Curran Associates, Inc., Vol. 25, 2012. [Online] : https://proceedings.neurips.cc/paper/2012/file/c399862d3b9d6b76c8436e924a68c45b-Paper.pdf", "doi": null, "title": null, "authors": null, "source": null, "year": null, "volume": null, "pages": null }, { "order": 16, "raw": "E. S. Gedraite, M. Hadad, \"Investigation on the effect of a Gaussian Blur in image filtering and segmentation,\" Proceedings ELMAR-2011, 14-16 Sept. 2011, pp. 393-396. ISBN: 978-1-61284-949-2", "doi": null, "title": null, "authors": null, "source": null, "year": null, "volume": null, "pages": null }, { "order": 17, "raw": "Gradio Internet sitesi [Online]: https://gradio.app", "doi": null, "title": null, "authors": null, "source": null, "year": null, "volume": null, "pages": null }, { "order": 18, "raw": "Hugging Face Internet sitesi [Online]: https://huggingface.co", "doi": null, "title": null, "authors": null, "source": null, "year": null, "volume": null, "pages": null }, { "order": 19, "raw": "ImgCERO Project İnternet Sitesi: https://huggingface.co/spaces/simayhosmeyve/Image_Enhancement", "doi": null, "title": null, "authors": null, "source": null, "year": null, "volume": null, "pages": null }, { "order": 20, "raw": "T. A. Möller, J. Nilsson, “Understanding SSIM”, 2020: arXiv Preprint: arXiv:2006.13846 [eess.IV] https://doi.org/10.48550/arXiv.2006.13846", "doi": null, "title": null, "authors": null, "source": null, "year": null, "volume": null, "pages": null }, { "order": 21, "raw": "O. Keleş, M. A. Yılmaz, A. M. Tekalp, C. Korkmaz, Z. Doğan, “On the Computation of PSNR for a Set of Images or Video”, 2021: arXiv Preprint : arXiv:2104.14868 [eess.IV] https://doi.org/10.48550/arXiv.2104.14868", "doi": null, "title": null, "authors": null, "source": null, "year": null, "volume": null, "pages": null }, { "order": 22, "raw": "A. Horé and D. Ziou, \"Image Quality Metrics: PSNR vs. SSIM,\" 2010 20th International Conference on Pattern Recognition, 2010, pp. 2366- 2369, doi: 10.1109/ICPR.2010.579.", "doi": null, "title": null, "authors": null, "source": null, "year": null, "volume": null, "pages": null }, { "order": 23, "raw": "E. van der Knaap, A. Ralescu, B. Ramsay, S. Visa, “Confusion Matrixbased Feature Selection”, MAICS , volume 710 of CEUR Workshop Proceedings, page 120-127.: http://ceur-ws.org/Vol-710/paper37.pdf", "doi": null, "title": null, "authors": null, "source": null, "year": null, "volume": null, "pages": null }, { "order": 24, "raw": "T. Nguyen, “Deep learning-based Face detection using the YOLOv3 algorithm” İnternet Sitesi [Online]: https://github.com/sthanhng/yoloface", "doi": null, "title": null, "authors": null, "source": null, "year": null, "volume": null, "pages": null } ], "statements": { "acknowledgements": null, "conflict_of_interest": null, "data_availability": null, "ai_disclosure": null }, "credit": [], "history": { "published": "2022-06-01" } }