Please use this identifier to cite or link to this item: http://hdl.handle.net/10609/90045
Title: Detección del carcinoma de glándula mamaria mediante termografía infrarroja usando redes neuronales profundas
Author: Fernández Ovies, Francisco Javier
Director: Prados Carrasco, Ferran  
Tutor: Alférez, Santiago  
Abstract: The main objective of this work is the evaluation of a convolutional neuronal network that facilitates the early detection of breast cancer using infrared thermography images. The literature shows satisfactory results in previous studies using other algorithms. The main challenges of using neural networks are two, having a large number of images and processing time. The approach is to use the high-level functions implemented in the fast.ai library, on the Pytorch platform, which have obtained excellent results in the image classification, as endorsed by their successes in the Kaggle competitions, a world reference in technologies Machine Learning. The images are organized in the three usual groups: training, validation and testing; what has told us to contrast the results using different pre-trained architectures (renet18, resnet34, resnet50, resnet152, vgg16 and vgg19) use the processing time, assess their classification by means of confusion matrices, obtain the best result with resnet34, with a accuracy in the 0.985 test.
Keywords: breast cancer
termography
deep neuronal network
Document type: info:eu-repo/semantics/masterThesis
Issue Date: 2-Jan-2019
Publication license: http://creativecommons.org/licenses/by-nc-nd/3.0/es/  
Appears in Collections:Bachelor thesis, research projects, etc.

Files in This Item:
File Description SizeFormat 
ffernandezovTFM0119memoria.pdfMemoria del TFM1,26 MBAdobe PDFThumbnail
View/Open