IJARP Impact Factor(2018): 4.908

International Journal of Advanced Research and Publications!

Paper Details: A Deep Neural Network Solution For Malignant Melanoma Detection

Volume 3 - Issue 11, November 2019 Edition
[Download Full Paper]

Bishwa Raj Poudel, Bibash Adhikari, Amrit Koirala, Sunil Dahal
Computer Aided Detection and Diagnosis, Machine Learning, Neural nets, Pattern Recognition, Segmentation.
This paper is an approach to predict the probability of an infected dermatoscopic image being cancerous (malignant melanoma) or not (benign) using learning power of Layered Convolutional Neural Network. Since the infected part of malignant highly resembles benign images, melanoma being the deadly fatal disease and confusion of dermatologists on differentiating malignant from benign easily has made this research a perfect application of machine learning to ease the field of Medical Science. Also, this research aims to propose computerized detection of malignant melanoma thereby preventing costly biopsy procedures that is otherwise done in clinic for diagnosing the disease. For this research, multiple architectures of layered neural network are designed, implemented and training is done by feeding those networks to around 13K dermatoscopic images using different training algorithms, multiple filters and tuning hyper-parameters in those algorithms to get the optimum result. 3 of the best performing milestone architectures with their hyper-parameter values and result analysis of those architectures is reflected in this paper. The best result we could achieve was with transfer learning using Google’s Inception-v3 which resulted out the staggering 94.02% of training accuracy and 89.73% of validation accuracy with f1-score of 0.81.
[1] M. Heath, N. Jaimes, B. Lemos, A. Mostaghimi, L. C. Wang, P. F. Peñas, and P. Nghiem, “Clinical characteristics of Merkel cell carcinoma at diagnosis in 195 patients: the AEIOU features,” Journal of the American Academy of Dermatology, vol. 58, no 3, pp. 375-381, 2008.

[2] American Cancer Society, “Facts & Figures 2019,” American Cancer Society, Atlanta, Ga. 2019.

[3] N. K. Mishra and M. E. Celebi, (2016, Jan). An Overview of Melanoma Detection in Dermoscopy Images using Image Processing and Machine Learning. [Online]. Available: https://arxiv.org/ftp/arxiv/papers/1601/1601.07843.pdf

[4] The American Cancer Society medical and editorial content team, (2019, Feb). Survival Rates for Melanoma Skin Cancer. [Online]. Available: https://www.cancer.org/cancer/melanoma-skin-cancer/detection-diagnosis-staging/survival-rates-for-melanoma-skin-cancer-by-stage.html

[5] A. Green, N. Martin, J. Pfitzner, M. O’Rourke, and N. Knight, “Computer Image Analysis in the Diagnosis of Melanoma,” Journal of the American Academy of Dermatology, pp. 958-964, 1994.

[6] A. Masood and A. A. Al-Jumaily, “Computer Aided Diagnostic Support System for Skin Cancer: A Review of Techniques and Algorithms,” International Journal of Biomedical Imaging, Article Id: 323268, Jul, 2013.

[7] J. F. Aitken, J. Pfitzner, D. Battistutta, P. K. O’Rourke, A. C. Green, and N. G. Martin, “Reliability of computer image analysis of pigmented skin lesions of Australian adolescents”, Cancer, vol. 78, no 2, pp. 252-257, Jul, 1996.

[8] Y. Chang, R. J. Stanley, R. H. Moss, and W. Van-Stoecker, “A systematic heuristic approach for feature selection for melanoma discrimination using clinical images,” Skin Research and Technology, vol. 11, no 3, pp. 165-178, Aug, 2005.

[9] Z. She, Y. Liu, and A. Damatoa, “Combination of features from skin pattern and ABCD analysis for lesion classification,” Skin Research and Technology, vol. 13, no 1, pp. 25-33, Feb, 2007.

[10] N. Fassihi, J. Shanbehzadeh, A. Sarafzadeh, E. Ghasemi, “Melanoma Diagnosis by the use of Wavelet Analysis based on Morphological Operators,” in Proc. IMECS, Hong Kong, 2011.

[11] International Skin Imaging Collaboration. ISIC Archive: The International Skin Imaging Collaboration: Melanoma Project. [Online]. Available: https://isic-archive.com/

[12] M. E. Celebi, Q. Wen, H. Iyatomi, K. Shimizu, H. Zhon, and G. Schaefer, “A State-of-the-Art Survey on Lesion Border Detection in Dermoscopy Images,” in Dermoscopy Image Analysis, M. E. Celebi, T. Mendonca, and J. S. Marques, Eds., Boca Raton, CRC Press, 2015, pp. 97-129.

[13] F. Bogo, F. Peruch, A. B. Fortina and E. Peserico, "Where’s the lesion? Variability in human and automated segmentation of dermoscopy images of melanocytic skin lesions," in Dermoscopy Image Analysis, M. E. Celebi, T. Mendonca and J. S. Marques, Eds., Boca Raton, CRC Press, 2015, pp. 67-96.

[14] Q. Abbas, I. F. Garcia, M. E. Celebi, and W. Ahmad, "A Feature‐Preserving Hair Removal Algorithm for Dermoscopy Images," Skin Research and Technology, vol. 19, no. 1, pp. 27-36, 2013.

[15] Q. Abbas, M. E. Celebi, and I. F. García, "Hair removal methods: a comparative study for dermoscopy images," Biomedical Signal Processing and Control, vol. 6, no. 4, pp. 395-404, 2011.

[16] R. Garnavi, M. Aldeen, M. E. Celebi, G. Varigos, and S. Finch, "Border detection in dermoscopy images using hybrid thresholding on optimized color channels," Computerized Medical Imaging and Graphics, vol. 35, no. 2, pp. 105-115, 2011.

[17] G. Schaefer, M. I. Rajab, M. E. Celebi, and H. Iyatomi, "Colour and contrast enhancement for improved skin lesion segmentation," Computerized Medical Imaging and Graphics, vol. 35, no. 2, pp. 99-104, 2011.

[18] R. J. Friedman, D. S. Rigel, and A. W. Kopf, "Early detection of malignant melanoma: The role of physician examination and self‐examination of the skin," CA: a cancer journal for clinicians, vol. 35, no. 3, pp. 130-151, 1985.