Deep Learning-based Type Identification of Volumetric MRI Sequences
Authors:
Jean Pablo Vieira de Mello,
Thiago M. Paixão,
Rodrigo Berriel,
Mauricio Reyes,
Claudine Badue,
Alberto F. De Souza,
Thiago Oliveira-Santos
Abstract:
The analysis of Magnetic Resonance Imaging (MRI) sequences enables clinical professionals to monitor the progression of a brain tumor. As the interest for automatizing brain volume MRI analysis increases, it becomes convenient to have each sequence well identified. However, the unstandardized naming of MRI sequences makes their identification difficult for automated systems, as well as makes it di…
▽ More
The analysis of Magnetic Resonance Imaging (MRI) sequences enables clinical professionals to monitor the progression of a brain tumor. As the interest for automatizing brain volume MRI analysis increases, it becomes convenient to have each sequence well identified. However, the unstandardized naming of MRI sequences makes their identification difficult for automated systems, as well as makes it difficult for researches to generate or use datasets for machine learning research. In the face of that, we propose a system for identifying types of brain MRI sequences based on deep learning. By training a Convolutional Neural Network (CNN) based on 18-layer ResNet architecture, our system can classify a volumetric brain MRI as a FLAIR, T1, T1c or T2 sequence, or whether it does not belong to any of these classes. The network was evaluated on publicly available datasets comprising both, pre-processed (BraTS dataset) and non-pre-processed (TCGA-GBM dataset), image types with diverse acquisition protocols, requiring only a few slices of the volume for training. Our system can classify among sequence types with an accuracy of 96.81%.
△ Less
Submitted 6 June, 2021;
originally announced June 2021.
BIOSTIRLING-4SKA : A cost effective and efficient approach for a new generation of solar dish-Stirling plants based on storage and hybridization; An Energy demo project for Large Scale Infrastructures",
Authors:
Domingos Barbosa,
Paulo Andre,
Paulo Antunes,
Tiago Paixao,
Carlos Marques,
Arnold van Ardenne,
Dion Kant,
Nico Ebbendorf,
Luis Saturnino Gonzalez,
Diego Rubio,
Lourdes Verdes-Montenegro,
Emilio Garcia,
Irene Calama,
Francisco Caballero,
Per Eskilson,
Lars Gustavsson,
Conny Anderson,
Jonah Lindh,
Reinhard Keller,
Manuel Silva Perez,
Gonzalo Lobo,
Valeriano Ruiz,
Javier Pino,
Luis Valverde,
Carsten Holze
, et al. (6 additional authors not shown)
Abstract:
The BIOSTIRLING - 4SKA (B4S) is a EU demonstration project dealing with the implementation of a cost-effective and efficient new generation of solar dish-Stirling plants based on hybridization and efficient storage at the industrial scale. The main goal of the B4S demonstration project is the generation of electric power using simultaneously solar power and gas to supply an isolated system and act…
▽ More
The BIOSTIRLING - 4SKA (B4S) is a EU demonstration project dealing with the implementation of a cost-effective and efficient new generation of solar dish-Stirling plants based on hybridization and efficient storage at the industrial scale. The main goal of the B4S demonstration project is the generation of electric power using simultaneously solar power and gas to supply an isolated system and act as a scalable example of potential power supply for many infrastructures, including future sustainable large scientific infrastructures. B4S build an interdisciplinary approach to address reliability, maintainability and costs of this technology. In April 2017, B4S successfully tested in Portugal the first world Stirling hybrid system providing about 4kW of power to a phased array of antennas, overcoming challenges in Stirling and hybridization and smartgrid technologies. B4SKA Consortium, with fourteen companies from six European countries, has performed the engineering, construction, assembly and experimental exploitation, under contract signed with the European to develop on off-grid demonstrator in Contenda (Moura) Portugal.
△ Less
Submitted 11 September, 2018; v1 submitted 8 December, 2017;
originally announced December 2017.
