Prof. João Luiz Carvalho - Research Projects


Ongoing research projects

Reconstruction of magnetic resonance flow imaging data using parallel imaging

Aims: to use parallel imaging reconstruction to reduce artifacts associated with temporal acceleration of spiral Fourier velocity encoded magnetic resonance imaging. Spiral FVE is a method for flow quantitation which may be particularly useful for estimating pressure gradients across stenoses in arteries and heart valves. Acquisitions may be accelerated by 18-fold using a combination of variable-density sampling, partial Fourier and temporal acceleration. We investigate the reduction of reconstruction artifacts due to temporal acceleration using multi-coil acquisitions.

Team: João L. A. Carvalho (principal investigator), Davi M. L. Leite (collaborator, University of Southern California), Gabriel S. Freitas (M.Sc. student)

Funding: MCT/CNPq/Edital Universal (Faixa A); CNPq/PIBIC (Scientific Initiation Fellowship to Davi M. L. Leite)

Parallelized reconstruction of magnetic resonance flow imaging data on multi-core processors and GPGPUs

Aims: to use GPU-based algorithms and multi-core processing to speed up reconstruction of multi-coil spiral FVE data. Spiral FVE is a method for MRI flow quantitation which may be particularly useful for estimating pressure gradients across stenoses in arteries and heart valves. Spiral FVE data are multidimensional - m(x,y,z,v,t) - and non-uniformly sampled in kx,ky. We investigate the use of parallel processing to speed-up the reconstruction of such datasets.

Team: João L. A. Carvalho (principal investigator), Thales H. Dantas (M.Sc. student), Gabriel S. Freitas (M.Sc. student), Rosana R. Lima (undergraduate student)

Funding: DPP/UnB Edital 10/2012 Apoio à Pesquisa de Novos Docentes;   CAPES/Edital Pró-equipamentos Institucional; CNPq/PIBIC (Scientific Initiation Fellowship to Rosana R. Lima)

Compressed sensing acceleration of spiral Fourier velocity encoded MRI

Aims: to reduce the duration of the spiral Fourier velocity encoding exam using compressed sensing. Compressed sensing is a technique that explores data sparsity for estimating signals and images from heavily-undersampled data. Spiral FVE datasets are multi-dimensional, and therefore highly suitable for compressed sensing acceleration.

Team: João L. A. Carvalho (advisor), Gabriel L. S. L. Oliveira (undergraduate student), Cristiano J. Miosso (collaborator)

Funding: CNPq/PIBIC (Scientific Initiation Fellowship to Gabriel L. S. L. Oliveira)

Segmentation of the aorta in real-time magnetic resonance image sequences

Aims: to improve the segmentation of the cross-section of the aorta in real-time MRI images. Real-time phase-contrast MRI is capable of measuring the stroke volume associated with each individual heartbeat. This requires accurate segmentation of the aorta in low resolution/low contrast images.

Team: João L. A. Carvalho (advisor), Gustavo M. Q. Mendonça (undergraduate student), Juliana F. Camapum (collaborator), Bruno L. Macchiavello (collaborator)

Past members: Thiago Z. Viana (undergraduate student), Gustavo M. Gondim (undergraduate student)

Publications:
- CBEB paper (coming soon)
- ISBI paper (coming soon)
- ISMRM paper (coming soon)
- Senior project: "Segmentation of the aorta in sequences of cardiac nuclear magnetic resonance images" (in Portuguese) [PDF]

Funding: CNPq/PIBIC (Scientific Initiation Fellowship to Gustavo M. Q. Mendonça)

Computational fluid dynamics simulation of carotid flow driven by Fourier velocity encoded magnetic resonance imaging

Aims: to use FVE-MRI data to improve the simulation of blood flow in the carotid bifurcation. Carotid atherosclerosis is the leading cause of thrombotic stroke. Serial and noninvasive assessment of carotid flow would be of value to better our understanding of the causal relationships between hemodynamics and the process of atherosclerotic plaque formation, growth, and rupture.

Publications:
- ISBI paper (coming soon)
- ISMRM paper (coming soon)

Team: João L. A. Carvalho (advisor), Vinícius C. Rispoli (doctoral student), Jon F. Nielsen (collaborator, University of Michigan)

Matlab software for detrended fluctuation analysis of heart rate variability

Aims: to design software for DFA analysis of heart rate variability. The analysis of HRV is an important tool for the assessment of the autonomic regulation of circulatory function. DFA is useful in the analysis of nonstationary HRV signals, since it involves removing fluctuation trends from the signal.

Team: João L. A. Carvalho (advisor), Daniel L. F. Almeida (undergraduate student)

Past members: Fernanda S. Leite (undergraduate student)

Publications:
- CBEB paper (coming soon)
- Senior project: "Development of algorithms for analysis of heart rate variability" (in Portuguese) [PDF]
- Conference paper: “Matlab software for detrended fluctuation analysis of heart rate variability” [PDF]

Estimation of the respiratory signal from real-time cardiac magnetic resonance images

Aims: to improve the estimation of the respiratory signal from real-time MRI images. Real-time phase-contrast MRI is capable of measuring the stroke volume associated with each individual heartbeat. In the future, we will investigate the relationship between stroke volume variability, heart rate variability and respiration under different sympathetic and parasympathetic stimuli.

Team: João L. A. Carvalho (advisor), Gustavo M. Q. Mendonça (undergraduate student)

Past members: Aline B. Alves (undergraduate student), Leila P. Morais (undergraduate student), Francisco Frantz (undergraduate student)

Analysis of surface electromoyography signals for evaluation of muscular fatigue

Aims: to use signal and image processing techniques for analysis of surface electromyopgraphic signals. The analysis of SEMG signals, including the frequency-domain analysis and estimation of conduction velocity, allows the evaluation of muscular fatigue. We investigate new acquisition protocols and signal processing techniques for SEMG analysis.

Team: Adson F. da Rocha (advisor), João L. A. Carvalho (co-advisor), Fabiano A. Soares (doctoral student), Gustavo M. Q. Mendonça (undergraduate student), Fabiano P. Schwartz (collaborator), Francisco A. O. Nascimento (collaborator). Marcelino M. Andrade (collaborator)

Past research projects

Neural-network classification of heart rate variability signals using multiple analysis techniques

Aims: to use neural networks to process coefficients obtained from multiple techniques for analysis of heart rate varaibility signals. The analysis of HRV is an important tool for the assessment of the autonomic regulation of circulatory function.

Team: João L. A. Carvalho (advisor), Rosana R. Lima (undergraduate student)

Publications:
- CBEB paper 1 (coming soon)
- CBEB paper 2 (coming soon)
- Senior project (cooming soon)

[Back to Home]