Hair Removal and Lesion Segmentation with FCN8-ResNetC and Image Processing in Images of Skin Cancer

Year 2022, , 231 - 238, 30.04.2022

Cihan Akyel , Nursal Arıcı

https://doi.org/10.17671/gazibtd.1060330

Cited By: 5

Abstract

Skin cancer is quite common. Early detection is crucial for the treatment of skin cancer. Methods based on computer technology (deep learning, image processing) are now increasingly used to diagnose skin cancer. These methods can eliminate human error in the diagnostic process. Removing hair noise from lesion images is essential for accurate segmentation. A correctly segmented lesion image increases the success rate in diagnosing skin cancer. In this study, a new FCN8-based approach for hair removal and segmentation in skin cancer images is presented. Higher success was achieved by adding ResNetC to FCN8. ResNetC is a new model based on ResNet. Two datasets were used for the study: ISIC 2018 and PH2. Training success was 89.380% for hair removal and 97.050% for lesion segmentation. 3000 hair masks were created as part of the study to remove hair noise in the lesion images.

Keywords

deep learning, hair removal, skin cancer, FCN, ResNet

Cilt Kanseri Görüntülerinde FCN8-ResNetC ve Görüntü İşleme ile Kıl Temizliği ve Lezyon Bölütleme

Abstract

Cilt kanseri oldukça yaygın görülmektedir. Cilt kanseri tedavisinde erken tespit önemlidir. Artık cilt kanseri tanısında bilgisayar teknolojisi temelli yöntemler (derin öğrenme, görüntü işleme) daha yaygın olarak kullanılmaktadır. Bu yöntemler ile tanı sürecinde insan hatası ortadan kaldırılabilir. Lezyon görüntüleri üzerindeki kıl gürültüsünün temizlenmesi doğru bölütleme için önem teşkil eder. Doğru bölütlenmiş lezyon görüntüsü ile cilt kanseri tanısında başarı oranı artacaktır. Bu çalışma, cilt kanseri görüntülerinde kıl temizliği ve bölütleme için FCN8 tabanlı yeni bir yaklaşım sunmaktadır. FCN8 algoritmasına ResNetC eklenerek başarı artışı sağlanmıştır. ResNetC ResNet tabanlı yeni bir modeldir. Çalışmada ISIC 2018’e ait iki veri seti ve PH2 veri seti kullanıldı. Kıl temizliğinde eğitim başarısı %89.380, lezyon bölütlemesinde ise %97.050 olarak elde edildi. Lezyon görüntülerindeki kıl gürültüsü temizliği için 3000 kıl maskesi çalışmada kapsamında oluşturulmuştur.

Keywords

derin öğrenme, kıl temizliği, cilt kanseri, FCN, ResNet

References

• O. Baykara, “Current Modalities in Treatment of Cancer”, Balıkesir Health Sciences Journal, 5(3), 154-165, 2016.
• Internet: WHO, https://www.who.int/news-room/fact-sheets/detail/cancer, 20.10.2021.
• R. L. Siegel, K.D. Miller KD, Jemal A. “Cancer statistics”, ACS Journal, 71(1), 7-33, 2021.
• H. M. Unver, E. Ayan, “Skin Lesion Segmentation in Dermoscopic Images with Combination of YOLO and GrabCut Algorithm”, Diagnostics Journal, 9(72), 1-21, 2019.
• K. H. Güngör, Metastaz Yapmamış Melanoma Ve Melanoma Dışı Deri Kanserleri İçin Geliştirilmiş Olan Deri Kanseri İlişkili Yaşam Kalitesi Ölçeğinin (Dkykö) Türkçe Geçerlilik Ve Güvenilirliğinin Araştırılması, Tıpta Uzmanlık Tezi, Ankara Üniversitesi Tıp Fakültesi, 2016.
• Internet: Ryerson University, https://rshare.library.ryerson.ca/articles/thesis/Skin_Lesion_Segmentation_Techniques_for_Melanoma_Diagnosis_Comparative_Studies/14649345/1, 18.01.2022.
• Internet: Arxiv, https://arxiv.org/ftp/arxiv/papers/1904/1904.11126.pdf, 25.02.2021.
• M. A. Kadampur, S. A. Riyaee, “Skin cancer detection: Applying a deep learning-based model-driven architecture in the cloud for classifying dermal cell images”, Informatics in Medicine Unlocked Journal, 18, 1-6, 2020.
• M. Senan, M. Jadhav, “Classification of Dermoscopy Images for Early Detection of Skin Cancer – A Review”, International Journal of Computer Applications, 178(17), 37-43, 2019.
• Internet: Science Direct, https://www.sciencedirect.com/science/article/pii/S1877050916305865, 11.05.2021.
• Z. Faisal, N. Abbadi, “New Segmentation Method for Skin Cancer Lesions”, Journal of Engineering and Applied Sciences, 12(21), 5598-5602, 2017.
• S. Jain, V. Jagtap, N. Pise, “Computer-aided Melanoma skin cancer detection using Image Processing”, Procedia Computer Science, 48, 735-740, 2015.
• T. Lee, V. Ng, R. Gallagher, A. Coldman, D. McLean, “Dullrazor: A Software Approach to Hair Removal from Images”, Computers in biology and medicine, 27(6), 533-543, 1997.
• H. El-Khatib, D. Popescu, L. Ichim, “Deep Learning-Based Methods for Automatic Diagnosis of Skin Lesions”, Sensors Journal, 20(6), 1-25, 2020.
• K. Zafar, S. O. Gilani, A. Waris, A. Ahmed, M. Jamil, M. A. Khan, A. S. Kaskif, “Skin Lesion Segmentation from Dermoscopic Images Using Convolutional Neural Network”, Sensors Journal 2020; 20(6). DOI: 10.3390/s20061601.
• Celebi, E.C., Aslandoğan, A.A., Stoecker WV, Iyatomi H, Oka H, et al. “Unsupervised Border Detection in Dermoscopy Image”, Skin Researchand Technology, 13(4), 454- 462, 2007.
• D. N. H. Thanh, N. H. Hai, P. Tiwari, H. L. Minh, “Skin lesion segmentation method for dermoscopic images with convolutional neural networks and semantic segmentation”, Computer Optics, 120, 121-129, 2021.
• C. Akyel, N. Arıcı, “A New Approach to Hair Noise Cleaning and Lesion Segmentation in Images of Skin Cancer”, Journal of Polytechnic, 23(3), 821-828, 2020.
• N. Şahin, N. Alpaslan, “SegNet Mimarisi Kullanılarak Cilt Lezyon Bölütleme Performansının İyileştirilmesi”, Avrupa Bilim ve Teknoloji Dergisi, Özel Sayı, 40-45, 2020.
• Brahmbhatt1, P., Rajan, S. N. “Skin Lesion Segmentation using SegNet with Binary CrossEntropy”, International Conference on Artificial Intelligence and Speech Technology (AIST2019), 14-15th November 2019.
• L. Talavera-Martínez, P. Bibiloni and M. González-Hidalgo, "Hair Segmentation and Removal in Dermoscopic Images Using Deep Learning", in IEEE Access, 9, 2694-2704, 2021.
• L. Wei, N.J.R. Alex, T. Tardi, Z. Zhemin, “Digital hair removal by deep learning for skin lesion segmentation”, Pattern Recognition,117, 1-15, 2021.
• K. Polat, A. S. Ashour, Y. Guo, E. Kucukkulahli, P. Erdogmus, “A hybrid dermoscopy images segmentation approach based on neutrosophic clustering and histogram estimation”, Applied Soft Computing 69, 426-434, 2018.
• Abdulhamid, M., Sahiner, A., Rahebi,J. “New Auxiliary Function with Properties in Nonsmooth Global Optimization for Melanoma Skin Cancer Segmentation”, Hindawi BioMed Research International, 1, 2020.
• K.Hasan, L. Dahal, P. N. Samarakoon, F. I. Tushara, R. Marti, “DSNet: Automatic Dermoscopic Skin Lesion Segmentation”, Computers in biology and medicine, 120, 426-434, 2020.
• Internet: Stanford University, https://web.stanford.edu/~kalouche/docs/Vision_Based_Classification_of_Skin_Cancer_using_Deep_Learning_(Kalouche).pdf, 03.01. 2021.
• C. Akyel, N. Arıcı, “LinkNet-B7: Noise Removal and Lesion Segmentation in Images of Skin Cancer”, Mathematics, 736-751, 2022.
• Internet: Task 3: LesionDiagnosis: Training, https://challenge2018.isicarchive.com/task3/training/, 20.10.2019.
• Internet: ISIC 2018: Skin Lesion Analysis Towards Melanoma Detection, https://challenge2018.isic-archive.com/, 15.10.2019.
• Internet: PH2 Dataset, https://www.fc.up.pt/addi/ph2%20database.html, 03.12.2021.
• Internet: Arxiv, https://arxiv.org/pdf/1411.4038.pdf, 10.05.2021.
• A. R. L´opez, S. Che, Skin Lesion Detection From Dermascopic Images Using Convolutional Neural Networks, A Degree Thesis, Polytechnic University of Catalonia, Barcelona, Spain, 2017.
• Internet: Softmax, https://towardsdatascience.com/additive-margin-softmax-loss-am-softmax- 912e11ce1c6b#:~:text=In%20short%2C%20Softmax%20Loss%20is,negative%20logarithm%20of%20the%20probabilities 20.03.2022.
• Internet: Keras, https://keras.io/api/optimizers/adam/25.12.2021.
• Y. Wang, A. Rahman, “Optimizing Intersection-Over-Union in Deep Neural Networks for Image Segmentation”, Conference: International Symposium on Visual Computing, 10 December 2016.
• T. Phan, S. Kim, H. Yang, G. Lee, “Skin Lesion Segmentation by U-Net with Adaptive Skip Connection and Structural Awareness”, Applied sciences, 11(4528), 1-14, 2021.
• F. Bagheri, M. J. Tarokh, M. Ziaratban, “Skin lesion segmentation based on mask RCNN, Multi Atrous Full-CNN, and a geodesic method”, Int J Imaging Syst Technol, 31(3), 1609–1624, 2021.
• C. K. Roy, J. R. Cordy, and R. Koschke. “Comparison and Evaluation of Code Clone Detection Techniques and Tools: A Qualitative Approach”, Sci. Comput. Program., 74(7), 470–495, 2009.