Melanoma Skin Cancer Detection Using Deep Learning

Notice

This is an unedited manuscript accepted for publication and provided as an Article in Press for early access at the author’s request. The article will undergo copyediting, typesetting, and galley proof review before final publication. Please be aware that errors may be identified during production that could affect the content. All legal disclaimers of the journal apply.

Year : 2025 | Volume : 14 | Issue : 02 | Page : 1-9
    By

    Anusiya M,

  • Aswathypriya M,

  • Dharshini P,

  • D. Prabakar,

  1. UG Scholar, Department of Computer Science and Engineering, Karpagam College of Engineering, Coimbatore, Tamil Nadu, India
  2. UG Scholar, Department of Computer Science and Engineering, Karpagam College of Engineering, Coimbatore, Tamil Nadu, India
  3. UG Scholar, Department of Computer Science and Engineering, Karpagam College of Engineering, Coimbatore, Tamil Nadu, India
  4. Professor & Head of the Department, Department of Computer Science and Engineering, Karpagam College of Engineering, Coimbatore, Tamil Nadu, India

Abstract

document.addEventListener(‘DOMContentLoaded’,function(){frmFrontForm.scrollToID(‘frm_container_abs_192354’);});Edit Abstract & Keyword

Melanoma, a fatal type of skin cancer, is a major global health concern. For better patient outcomes, early and precise detection is essential. A branch of artificial intelligence called deep learning has demonstrated encouraging outcomes in medical image analysis, particularly the identification of skin cancer, in recent years. We present a new method for detecting melanoma skin cancer in this paper by utilizing the ResNet-50 architecture, a deep convolutional neural network (CNN). A pre-trained ResNet-50 model, which was first trained on a sizable dataset of varied images, is used in the suggested strategy to take advantage of the capability of transfer learning. This greatly reduces training time and boosts overall performance by allowing the network to learn low-level characteristics efficiently. We use a dataset of hundreds of images of skin lesions, both benign and malignant, to fine-tune the ResNet-50 architecture for melanoma classification precisely. To adapt the pretrained ResNet-50 to the binary classification problem, the final fully connected layers must be modified during the training process. We use the Adam optimizer during training and binary cross-entropy as the loss function. Furthermore, data augmentation techniques are used to decrease overfitting and boost dataset heterogeneity.

Keywords: Image segmentation, deep learning, convolutional neural network, Skin Cancer, Global Health

[This article belongs to Research and Reviews: Journal of Oncology and Hematology ]

How to cite this article:
Anusiya M, Aswathypriya M, Dharshini P, D. Prabakar. Melanoma Skin Cancer Detection Using Deep Learning. Research and Reviews: Journal of Oncology and Hematology. 2025; 14(02):1-9.
How to cite this URL:
Anusiya M, Aswathypriya M, Dharshini P, D. Prabakar. Melanoma Skin Cancer Detection Using Deep Learning. Research and Reviews: Journal of Oncology and Hematology. 2025; 14(02):1-9. Available from: https://journals.stmjournals.com/rrjooh/article=2025/view=0


document.addEventListener(‘DOMContentLoaded’,function(){frmFrontForm.scrollToID(‘frm_container_ref_192354’);});Edit

References

  1. Agarap AF. Deep Learning using Rectified Linear Units (Relu). Ar Xiv preprint. 2018. Available from: http://arxiv.org/abs/1803.08375.
  2. Khazaei Z, Ghorat F, Jarrahi AM, Adineh HA, Sohrabivafa M, Goodarzi EJWCRJ. Global incidence and mortality of skin cancer by histological subtype and its relationship with the human development index (HDI); An ecology study in 2018. World Cancer Res J. 2019;6(2):e13.
  3. An ecology study examined the association between the Human Development Index (HDI) and the incidence and death of skin cancer worldwide by histological subtype. World Cancer Res J. 2019;6:13.
  4. Cancer Council Australia. Skin cancer facts and figures [Internet]. 2010 [cited year unknown]. Available from: http://www.cancer.org.au/cancersmartlifestyle/SunSmart/Skincancerfactsandfigures.htm.
  5. Filler T, Bas P, Pevny T. The nuances of setting up BOSS: A breakdown of our steganographic system. In: International Workshop on the Hide of Information. Berlin: Springer; 2011. p. 59–70. doi:10.1007/978-3-642-24178-9_5.
  6. Schulman J, Sutskever I, Houthooft R, Chen X, Duan Y, Abbeel P. Info GAN: Interpretable representation learning using generative adversarial nets. In: Advances in Neural Information Processing Systems. 2016;29. Available from: https://proceedings.neurips.cc/paper/2016/file/7c9d0b1f96aebd7b5eca8c3edaa19ebb-Paper.pdf.
  7. Chen X, Huang Y, Kong C, Jiang K, Li Y, Pan J, et al. Dual contrastive learning for deraining deep images without pairs. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR); 2022. p. 2007–2016. doi:10.1109/CVPR52688.2022.00206.
  8. Deb K, Pratap A, Agarwal S, Meyarivan T. A fast and elitist multiobjective genetic algorithm: NSGA-II. IEEE Trans Evol Comput. 2002;6(2):182–197. doi:10.1109/4235.996017.
  9. El-Khalil R, Keromytis AD. Hydan: Program binaries conceal data. In: Lopez J, Qing. S, Okamoto E, editors. Information and Communications Security. Berlin: Springer; 2004. p. 187–199.
  10. Fard AM, Varasteh-A F, Akbarzadeh-T MR. A novel way to secure JPEG steganography using genetic algorithms. In: IEEE International Conference on Engineering of Intelligent Systems; 2006. p. 1–6. doi:10.1109/ICEIS.2006.1703168.
  11. Finlayson SG, Ito J, Zittrain JL, Beam AL, Kohane IS, Bowers JD. Adversarial attacks on medical machine learning. Science. 2019;363(6433):1287–1289. doi:10.1126/science. Aaw 4399.
  12. Fridrich J, Kodovsky J. Rich models for steganalysis of digital images. IEEE Trans Inf Forensics Secur. 2012;7(3):868–882. doi:10.1109/TIFS.2012.2190402.
  13. Geetha S, Kamaraj N. Enhancing picture steganalysis by employing MBEGA for feature selection. Ar Xiv preprint. 2010. arXiv:1008.2824. doi:10.48550/arXiv.1008.2824.
  14. Ghasemi E, Fassihi N, Shanbehzadeh J. Combining evolutionary algorithms and wavelet transforms for high capacity image steganography. In: Proceedings of the 2012 World Congress on Engineering; London, UK. 2012 Jul 4–6; 2188. p. 495–498. Available from: http://www.iaeng.org/publication/IMECS2011.
  15. Fierrez J, Camacho D, Martín A, Huertas-Tato J. Enhancing image classification by combining statistical indicators with CNNs. Inf Fusion. 2022;79:174–187. doi: 10.1016/j.inffus.2021.09.012.
  16. Wuhab AWA, Idris YIB, Ho AT, Hussain M, Jung KH. A survey of spatial image steganography. Signal Process Image Commun. 2018;65:46–66. doi: 10.1016/j.image.2018.03.012.
  17. Premaratne P, Vial PJ, Kadhim IJ, Halloran B. An extensive overview of image steganography, including methods, assessments, and directions for further study. Neurocomputing. 2019;335:299–326. doi: 10.1016/j.neucom.2018.06.075.
  18. Mandal PC, Paul G, Mukherjee I, Chatterji B. Steganography of digital images: A review of the literature. Inf Sci. 2022;609:1451–1488. doi: 10.1016/j.ins.2022.07.120.
  19. Martín A, Camacho D, Fuentes-Hurtado F, Naranjo V. Developing designs for deep neural networks to classify Android malware. In: Proceedings of the IEEE Congress on Evolutionary Computation (CEC); 2017. p. 1659–1666. doi:10.1109/CEC.2017.7969501.
  20. Khodaei M, Faez K. Employing LSB substitution and a genetic algorithm to hide images. In: Meniere J, Mammas D, Norbord F, Lazore O, Elmoataz A, editors. Image and Signal Processing. Berlin: Springer; 2010. p. 404–411. doi:10.1007/978-3-642-13681-8_47.
Regular Issue Subscription Review Article
Volume 14
Issue 02
Received 26/03/2025
Accepted 02/04/2025
Published 11/06/2025
Publication Time 77 Days
89

[first_name] [last_name]

My IP

PlumX Metrics