Implementación de Redes Generativas Adversarias (GANs) para la generación de imágenes de tejido humano

This project is based on the generation of images of human tissue, specifically, skin lesions such as melanoma or seborrheic keratosis, using a Generative Adversarial Network and a dataset of 2000 real images. The main application of the project in the field of medicine is to increase the global dat...

Descripción completa

Detalles Bibliográficos
Autor: Segura Cabrera, Pedro Juan
Tipo de recurso: tesis de maestría
Fecha de publicación:2019
País:España
Institución:Universitat Oberta de Catalunya (UOC)
Repositorio:O2, repositorio institucional de la UOC
OAI Identifier:oai:openaccess.uoc.edu:10609/90566
Acceso en línea:http://hdl.handle.net/10609/90566
Access Level:acceso abierto
Palabra clave:disciminator
GAN
deep learning
CNN
MVP
generator
artificial intelligence
discriminador
aprenentatge profund
generador
intel·ligència artificial
aprendizaje profundo
inteligencia artificial
Artificial intelligence -- TFM
Intel·ligència artificial -- TFM
Inteligencia artificial -- TFM
Descripción
Sumario:This project is based on the generation of images of human tissue, specifically, skin lesions such as melanoma or seborrheic keratosis, using a Generative Adversarial Network and a dataset of 2000 real images. The main application of the project in the field of medicine is to increase the global database of images of human tissue and thus, contribute to medical studies in the dermatological area. Images with a resolution of 64x64 have been obtained and validated through a representation of reduced dimensions (PCA and tSNE) of the whole set of images, real and generated ones.