INVESTIGADOR(ES) PRINCIPAL(ES):

Alvaro Angel Orozco Gutiérrez

8 horas

CODIGO CIE

6-16-1

Julián David Echeverry Correa

Coinvestigador

8 Horas

Hernán Felipe García Arias

Estudiante

0 Horas

Hernán Darío Vargas Cardona

Estudiante

0 Horas

Mauricio Alexander Alvarez Lopez

Coinvestigador

8 Horas

Julián Gil González

Estudiante

0 Horas

TIPO DE CONVOCATORIA

2015. Novena Convocatoria

TIPO DE PROYECTO

Investigación Aplicada

OBJETIVO(S)

General objetive: To develop a methodology based on stochastic processes for enhancing spatial resolution of medical imaging applied on segmentation of nerves, brain structures and tractography reconstruction. Specific objectives: - To develop a stage for pre-processing of medical images through adaptive filtering methods - To model scalar imaging (MRI and Ultrasound) with Gaussian processes and tensorial imaging (dMRI) employing Generalized Wishart processes for enhancing spatial resolution. - To optimize parameters in each model applying Markov Chain Monte-Carlo methods. - To validate the proposed methodology through error metrics for segmentation of brain structures, nerves recognition and tractography reconstruction.

RESUMEN

* Brain image smoothing improves the estimation of the diffusion tensors (DTs). The methods can reduce the noise level but they may introduce unwanted blurring if we select a wrong set of hyperparameters. * We modeled a Gaussian process regression (GPR) for 2D multi-slice images. Results achieved with our proposal outperform to B-spline interpolation, nearest-neighbor. Our methodology performs better in both super-resolution images and morphological validation. * We introduced a methodology for spatial resolution enhancement in ultrasound images based on a novel area in supervised learning known as learning from multiple annotators. Results achieved with the proposed methodology outperform to the other GPR methods. * We developed a probabilistic methodology to interpolate Diffusion Tensor Imaging (DTI) data. We model a DTI field as a Generalized Wishart process (GWP). Results obtained with GWP in synthetic and real DTI data outperform to commonly used geometric methods. Also, our proposed method guarantees positive definite tensors, excellent accuracy and it avoids an issue in tensorial interpolation known as swelling effect.

ESTADO

Concluido

FECHA DE INICIO

18/01/2016

FECHA DE FINALIZACION

18/07/2017

PRODUCTOS

Comparison of Preprocessing Methods for Diffusion Tensor Estimation in Brain Imaging

Artículos en revista A1 ó A2

EN PROCESO: A methodology for peripheral nerve segmentation using a multiple annotators approach based on Centered Kernel Alignment

Maestría o Especialidad clínica

EN PROCESO: Procesos generalizados de Wishart no Estacionarios para la Interpolación de Campos Tensoriales en Imágenes de Resonancia Magnética de Difusión

Maestría o Especialidad clínica

Spatial resolution enhancement in ultrasound images from multiple annotators knowledge

Artículos en revista A1 ó A2

Technical Report - Gaussian processes for slice-based super-resolution MR images

Artículos en revista A1 ó A2

Technical Report - Generalized Wishart processes for interpolation over diffusion tensor fields

Artículos en revista A1 ó A2