Please use this identifier to cite or link to this item:
http://hdl.handle.net/10609/82129
Title: | Servicio web para la clasificación de cáncer de piel usando redes neuronales |
Author: | Iglesias Garcia, Endika |
Tutor: | Alférez, Santiago |
Others: | Universitat Oberta de Catalunya Morán Moreno, Jose Antonio |
Abstract: | Skin cancer, the most common human malignancy, is primarily diagnosed visually, beginning with an initial clinical screening and followed potentially by dermoscopic analysis, a biopsy and histopathological examination. Automated classification of skin lesions using images is a challenging task due to the variability in appearance. Convolutional Neural Networks have a potential for such general and highly variable tasks. For this purpose, an accessible web platform will be developed, where users will send images of skin areas and the possible carcinogenic zones found will be analyzed and identified. This analysis will be performed by a deep convolutional neural network. We will use the ISIC dataset containing images classified as benign and malignant will be used, as well as other classifications such as the type of lesion presented in the image. Initially, a development of an own neuronal model using convolutional layers will be performed. Pre-trained networks such as InceptionV3, VGG16 and VGG19 will also be used. A preliminary analysis will be carried out, where the accuracy of each of these neural networks can be observed. The accuracy of each of these will then be evaluated to select the appropriate one for fine tuning and use in the web service. |
Keywords: | skin cancer web applications neural networks |
Document type: | info:eu-repo/semantics/masterThesis |
Issue Date: | Jun-2018 |
Publication license: | http://creativecommons.org/licenses/by-nc-nd/3.0/es/ |
Appears in Collections: | Trabajos finales de carrera, trabajos de investigación, etc. |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
endika2TFM0618memoria.pdf | Memoria del TFM | 20,31 MB | Adobe PDF | View/Open |
Share:
This item is licensed under a Creative Commons License