M
Miguel Ángel González Ballester
Researcher at Pompeu Fabra University
Publications - 218
Citations - 4320
Miguel Ángel González Ballester is an academic researcher from Pompeu Fabra University. The author has contributed to research in topics: Computer science & Segmentation. The author has an hindex of 25, co-authored 194 publications receiving 2913 citations. Previous affiliations of Miguel Ángel González Ballester include T-Systems & Catalan Institution for Research and Advanced Studies.
Papers
More filters
Posted Content
Balanced-MixUp for Highly Imbalanced Medical Image Classification
TL;DR: Balanced-MixUp as discussed by the authors proposes a method for sampling training data based on the popular MixUp regularization technique, which is referred to as balanced-mixup, which simultaneously performs regular and balanced sampling of the training data.
Journal ArticleDOI
Anatomical medial surfaces with efficient resolution of branches singularities.
TL;DR: Algebraic geometry concepts are used in an energy-based approach to compute a medial surface presenting a stable branching topology and an efficient GPU-CPU implementation using standard image processing tools are presented.
Journal ArticleDOI
Semi-Supervised Placental Vessel Segmentation from Fetoscopy Videos
Blanca Zufiria,Aregawi Halefom,Rodrigo Cilla,Mario Ceresa,Elisenda Bonet-Carne,Elisenda Eixarch,Miguel Ángel González Ballester,Ivan Macia,Karen López-Linares +8 more
Pulmonary Nodule Malignancy Classification Using its Temporal Evolution with Two-Stream 3D Convolutional Neural Networks.
Xavier Rafael-Palou,Anton Aubanell,Ilaria Bonavita,Mario Ceresa,Gemma Piella,Vicent J. Ribas,Miguel Ángel González Ballester +6 more
TL;DR: In this article, a two-stream 3D convolutional neural network was proposed to predict malignancy by jointly analyzing two pulmonary nodule volumes from the same patient taken at different time-points.