Please use this identifier to cite or link to this item: http://hdl.handle.net/10609/139006
Title: Algoritmo de clasificación de lesiones en exámenes mamográficos
Author: Bustos Pelegri, Joel
Tutor: Martínez Maldonado, Sergi
Others: Rius, Àngels  
Abstract: The use of computer vision techniques applied in the field of medicine allows us to accelerate the process of detection of any type of disease, helping specialists to carry out diagnoses and reducing the mortality rate when detecting possible symptoms during premature stages. Specifically, convolutional neural networks are part of the state-of-the-art in image classification tasks thanks to the fact that their two-dimensional architecture resembles the structure of the input data. In this work, 4 convolutional neural network architectures have been used to classify the different types of lesions present in mammographic images, as malignant or benign. The decisions made by each architecture have been combined using a Random Forest algorithm in order to emulate the diagnosis made by different specialists when analyzing a mammographic examination. The final tool generated from the sequential combination of classifiers has presented metrics of 92% for the classification of benign and malignant samples from the MIAS data set.
Keywords: breast cancer diagnosis
convolutional networks
Document type: info:eu-repo/semantics/masterThesis
Issue Date: 2-Jan-2022
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 
jbustospelegriTFM0122memoria.pdfMemoria del TFM4,66 MBAdobe PDFThumbnail
View/Open