Improving Breast Cancer Detection Using DCGANs Generated Synthetic Mammograms
Autor:
Garcia-Tiscar, Luis
Fecha:
11/09/2021Palabra clave:
Tipo de Ítem:
masterThesis
Resumen:
En el campo del Aprendizaje Automático, existe un compromiso entre la cantidad de datos con los que se entrena un modelo y los resultados del mismo. Uno de los principales problemas de la adquisición de datos son las anomalías, los valores erróneos o, principalmente, faltantes. En aplicaciones médicas, como el análisis de mamografías para la detección de tumores, estas anomalías pueden mermar el rendimiento del modelo y repercutir directamente en la vida de los pacientes, debido a un diagnóstico erróneo. Para subsanar este problema, presentamos un desarrollo de software capaz de generar mamografías sintéticas de alta resolución con presencia de tumores benignos y malignos a través de redes neuronales DCGAN (Deep Convolutional Generative Adversarial Networks). El objetivo es comprobar si estas imágenes sintéticas son clasificadas correctamente como tumores benignos o malignos por los clasificadores de cáncer de mama del estado del arte.
Descripción:
The amount of available training data and a Machine Learning model’s results are strongly related. Missing values are one of the main issues in this field, making models lose accuracy and overall quality in their results. Machine Learning models, especially in healthcare applications, need to be very accurate, because poor results lead to misdiagnosis and can affect patients' health. In order to solve this issue, we present an unsupervised ML model to generate high resolution synthetic mammograms based on DCGAN (Deep Convolutional Generative Adversarial Networks). Our goal is to create synthetic mammograms, indistinguishable from real ones, to feed these synthetic images to state-of-the-art breast cancer detection models and check if they are treated as real mammograms.
Ficheros en el ítem
Este ítem aparece en la(s) siguiente(s) colección(es)
Estadísticas de uso
| Año |
| 2012 |
| 2013 |
| 2014 |
| 2015 |
| 2016 |
| 2017 |
| 2018 |
| 2019 |
| 2020 |
| 2021 |
| 2022 |
| 2023 |
| 2024 |
| 2025 |
| Vistas |
| 0 |
| 0 |
| 0 |
| 0 |
| 0 |
| 0 |
| 0 |
| 0 |
| 0 |
| 0 |
| 93 |
| 63 |
| 89 |
| 94 |
| Descargas |
| 0 |
| 0 |
| 0 |
| 0 |
| 0 |
| 0 |
| 0 |
| 0 |
| 0 |
| 0 |
| 57 |
| 36 |
| 35 |
| 56 |
Ítems relacionados
Mostrando ítems relacionados por Título, autor o materia.
-
Time series analysis for COMEX platinum spot price forecasting using SVM, MARS, MLP, VARMA and ARIMA models: A case study
Menéndez-García, Luis Alfonso; García-Nieto, Paulino José; García-Gonzalo, Esperanza; Sánchez Lasheras, Fernando (Resources Policy, 2024)This article looks at predicting the price of platinum, along with 12 other commodity prices, using both time series and machine learning models. Platinum, characterised by its rarity and significant industrial and artistic ... -
Prevalence of burnout in paediatric nurses: A systematic review and meta-analysis
Pradas-Hernandez, Laura; Ariza, Tania ; Gómez-Urquiza, José Luis; Albendín-García, Luis; De la Fuente, Emilia I. ; Cañadas-De la Fuente, Guillermo A. (Plos One, 25/04/2018)Introduction Although burnout in paediatric nurses has been addressed in previous research, the heterogeneous nature of the results obtained and of the variables studied highlights the need for a detailed analysis of the ... -
Neuromuscular Fatigue and Metabolic Stress during the 15 Minutes of Rest after Carrying Out a Bench Press Exercise Protocol
Hernández-Lougedo, Juan; Heredia-Elvar, Juan Ramón; Maicas-Pérez, Luis; Cañuelo-Márquez, Ana María; Rozalén-Bustín, Manuel; de Jesús-Franco, Fernando; Garnacho-Castaño, Manuel V.; García-Fernández, Pablo; Maté-Muñoz, José Luis (Biology, 2022)Background: Velocity loss (VL) at 1 m·s−1 can help to determine neuromuscular fatigue after performing an exercise protocol. The aim of this study was to analyse muscle fatigue and metabolic stress during the 15 min that ...





