Towards to Deep Neural Network Application with Limited Training Data: Synthesis of Melanoma's Diffuse Reflectance Spectral Images
Proceedings of SPIE
The goal of our study is to train artificial neural networks (ANN) using multispectral images of melanoma. Since the number of multispectral images of melanomas is limited, we offer to synthesize them from multispectral images of benign skin lesions. We used the previously created melanoma diagnostic criterion p'. This criterion is calculated from multispectral images of skin lesions captured under 526nm, 663nm, and 964nm LED illumination. We synthesize these three images from multispectral images of nevus so that the p' map matches the melanoma criteria (the values in the lesion area is >1, respectively). Demonstrated results show that by transforming multispectral images of benign nevus is possible to get a reliable multispectral images of melanoma usable for ANN training.
artificial neural networks, diffuse reflectance, early diagnostics, skin cancer
Boločko, K., Bļizņuks, D., Uteshev, D., Lihacova, I., Lihachev, A., Čižovs, J., Bondarenko, A. Towards to Deep Neural Network Application with Limited Training Data: Synthesis of Melanoma's Diffuse Reflectance Spectral Images. In: Proceedings of SPIE, Germany, Minhene, 23-25 June, 2019. -: SPIE, 2019, pp.1-6. ISSN 1605-7422. e-ISSN 2410-9045. Available from: doi:10.1117/12.2527173