Aleksandra Dzieniszewska
supervisor: Ryszard Piramidowicz
Melanoma is one of the deadliest kinds of skin cancer, but it might be completely cured if detected early. Therefore, monitoring and early diagnosis of skin lesions are crucial in preventing cancer diseases. The major problem and challenge, however, is patients’ resistance to being diagnosed with cancer and limited access to specialists.
The automation of the diagnosis process might, to some extent, help in solving these problems by providing easy access to mass and sufficiently reliable testing methods, which patients might have access to in a family doctor’s office, or at home.
This work is focused on developing a new solution for a comprehensive automated skin analysis system allowing the classification of changes based on multimodal image data and deep learning models.
In particular, the application of neural networks to enhance the recognition of human skin lesions based on dermatoscopy images in RGB colorspace and with additional polarisation information was explored. The main goal of the research was to employ deep learning techniques to support the diagnosis process and provide full automatization. Multiple approaches to skin lesion diagnosis were tested to find the best accuracy using a limited data set and provided insight into diagnosis. The work is focused on different approaches to autoencoder neural networks that can be modified to perform reconstruction or segmentation and classification simultaneously.