MICLab - Medical Image Computing Lab

Welcome to the Medical Image Computing Lab (MICLab) website. We are a research group in computational techniques for image processing, specially for medical imaging.

MICLab is part of the School of Electrical and Computer Engineering (FEEC) at the State University of Campinas (UNICAMP), Brazil. We have joined projects with the Institute of Computing, the Medical Sciences Faculty (both at UNICAMP) and others. Our research has been supported by CAPES, CNPQ and FAPESP.

Opportunities in our group

We are currently looking for motivated undergraduate students to join our group and take part of one of our research projects:

For more information on each project:

Students will be encouraged to submit their project to PIBIC scholarships program (Programa Institucional de Bolsas de Iniciação Científica). The deadline is April 20th 2015.

People

Principal Investigators:

Roberto A. Lotufo

/media/Attachments/courseIA898A1S2015/homepage_lotufo/lotufo_2014.jpg

Professor Roberto A. Lotufo is a full professor at the School of Electrical and Computer Engineering, University of Campinas (UNICAMP). His principal interests are in the areas of Image Processing and Analysis, Mathematical Morphology, Image Segmentation and Medical Imaging.

Letícia Rittner

/media/Attachments/MICLab/MainPage/photo_lrittner.jpg

Professor Leticia Rittner is an Assistant Professor at the School of Electrical and Computer Engineering, University of Campinas (UNICAMP). Her principal interests are Medical Imaging, Image Processing and Analysis, Image Segmentation and Pattern Recognition

PhD students:

André Costa

/media/Attachments/homeandre/MainPage/foto_perfil.jpg

André Costa is a Ph.D. student working on the human corpus callosum characterization from magnetic ressonance images project. His main interests are:

  • graph theory
  • normalized cut
  • image segmentation and classification
  • pattern recognition
  • medical imaging

Mariana Pinheiro

/media/Attachments/courseIA369O1S2011/mar_h/eu.png

Mariana Pinheiro is a Ph.D. student working on the Analysis of White Matter Lesions (hyperintensity) in MRI. Her principal interests are:

  • Medical Imaging
  • Image Processing
  • Pattern Recognition
  • Image classification

Roberto Souza

/media/Attachments/courseTopics2S2013/robertomedeirossouzagmailcom/rob.png

Roberto Souza is a Ph.D. student working under orientation of Prof. Roberto de Alencar Lotufo on the area of Image Processing. His principal interests are:

  • Mathematical Morphology
  • Image Processing
  • Pattern Recognition
  • Image classification
  • CBIR
  • Digital forensics

Irene Fantini

/media/Attachments/courseIA8981S2015/homepage_irenefantini/Irene.jpg

Irene Fantini is a Ph.D. student working on the analysis of MRI quality using deep learning. Her principal interests are:

  • Medical Imaging
  • Image Processing
  • Image classification
  • Deep learning

Master students:

Danilo Pereira

/media/Attachments/MICLab/MainPage/Danilo.jpg

Danilo Pereira is a Master student working on the visualization and analysis of multi-voxel Magnetic Resonance Spectroscopy (MRS) of the brain. His principal interests are:

  • Magnetic Resonance Spectroscopy (MRS)
  • Medical Imaging
  • Image Processing

Silvia Collela

/media/Attachments/homesilviacolella/MainPage/photo_silviacolella.jpg

Silvia Colella is a Master student working with CT images of the lungs. Her principal interests are:

  • Computed Tomography (CT)
  • Medical Imaging
  • Image Processing
  • Image classification

William García

/media/Attachments/courseIA8981S2015/homepage_WilliamGarcia/WG_CV.jpg

William García is a Master student working on Image Processing and Machine Learning. His principal interests are:

  • Medical Imaging
  • Image Processing
  • e-Health
  • Machine learning

Gustavo Retuci Pinheiro

/media/Attachments/MICLab/MainPage/gustavo.jpg

Gustavo Retuci Pinheiro is a Master student working on Image Processing. His main interests are:

  • Medical Imaging
  • Image Processing
  • Pattern Recognition
  • DTI

Guilherme Saraiva Soares

/media/Attachments/courseIA8981S2015/homepage_guilhermesaraiva/Guilherme.png

Guilherme Saraiva Soares is a Master student working on techniques of people counting using deep learning. His main interests are:

  • Image processing
  • Pattern recognition
  • Video content analytics
  • Deep learning

Giovana Cover

/media/Attachments/MICLab/MainPage/gio_2.png

Giovana Cover is a Master student working on brain Image Processing. Her principal interests are:

  • Medical Imaging
  • Image Processing
  • Pattern Recognition
  • Image Classification
  • Neural Networks
  • Diffusion Tensor Imaging

Lívia Rodrigues

/media/Attachments/MICLab/MainPage/Livia.jpg

Lívia Rodrigues is a Master student working on Image Processing. Her principal interests are:

  • Medical Imaging
  • Image Processing
  • Pattern Recognition
  • Machine Learning

Former students

Andre Korbes

Pedro Ferro Freitas

Victor Oliveira

Research main topics

Diffusion Tensor Imaging (DTI) and Morphological Processing

/media/Attachments/DTI/MainPage/original.png

Figure 14: Diffusion Tensor Imaging (DTI)

Diffusion tensor imaging (DTI) is a relatively new modality of Magnetic Resonance Imaging (MRI) able to quantify the anisotropic diffusion of water molecules in highly structured biological tissues. It is unique in its ability to quantify changes in neural tissue microstructure within the human brain non-invasively.

Among the several DTI publications, we highlight two areas: segmentation and tractography. DTI-based segmentation aims to delineate regions with similar diffusion characteristics and is a necessary step for performing subsequent quantitative analysis and qualitative visualization. Another important application of DTI is the fiber tracking in the brain, which, in combination with functional MRI, might open a window on the important issue of brain connectivity.

Our goal is to develop methods and tools to process, analyze and visualize DTI using operators and concepts of mathematic morphology. See more

Human Corpus Callosum characterization from magnetic ressonance images

/media/Attachments/homeandre/project_abstract/curvas.png

Figure 15: Curves of scalar measurements extracted from DTI. 200 points along the midsagittal middle axis (Freitas, 2012).

The human corpus callosum (CC) is the largest white matter structure, composed mostly by neuronal fibers. Several studies associate CC abnormalities to a variety of brain illnesses. Those studies rely on measurements of the CC macro and micro structures on magnetic ressonance images. Nevertheless, only a few recent works (Park et al.,2011; Rittner, 2014) relates the CC microstructure to its extent. Figure 1 show curves of DTI scalar measurements along the midsagittal middle axis.

In this project we aim at the extraction of a CC signature from magnetic ressonance images, that could be useful for monitoring changes or to identify abnormalities. Other goals include the signature visualization, and the improvement of DTI spatial resolution by super-resolution techniques.

Publications

Analysis of white matter hyperintensity in brain MRI

/media/Attachments/homemariana/reuniao_cepid/samples.png

Figure 16: White Matter Hyperintensity

The brain white matter is responsible for the transmission of electrical signals through the central nervous system. Lesions in the brain white matter, called white matter hyperintensity (WMH), can cause a significant functional deficit. WMH are commonly seen in normal aging, but also in a number of neurological and psychiatric disorders. We propose here an automatic method for WHM analysis in order to distinguish regions of interest between normal and non-normal white matter (identification task) and also to distinguish different types of lesions based on their etiology: demyelinating or ischemic (classification task). The method combines texture analysis with the use of classifiers, such as Support Vector Machine (SVM), Nearst Neighboor (1NN), Linear Discriminant Analysis (LDA) and Optimum Path Forest (OPF). Experiments with real brain MRI data showed that the proposed method is suitable to identify and classify the brain lesions.

The watershed transform

/media/Attachments/watershed/MainPage/ws-example-1.jpg

Figure 17: The watershed transform

Introduced by Digabel and Lantuéjoul, the watershed transform was later used by Beucher and Lantuéjoul for contour detection. These introductory works, along with Meyer and Beucher work on morphological segmentation settled the concept of image segmentation based on a gradient image, where the grey levels are altitudes forming a surface with catchment basins submitted to a flooding process. When two basins touch, a barrier (a watershed line) is raised. Intuitively, these segmentation lines are points where a drop of water may slide to two different regional minima.

The goal of this project is to keep a place where the watershed transform may be discussed, providing different implementations in Python and also in C++ with CUDA. See more

The Optimum-Path Forest Classifier (OPF)

/media/Attachments/main/MainPage/opf.jpg

Classification is one of the most important tasks in machine learning, with a wide range of applications: computer vision, data mining, multimedia indexing, among others. There are several well-known classifiers, but only some of them are able to deal efficiently with multiclass problems. The Optimum-Path Forest (OPF) is a framework for supervised and unsupervised classifiers based on the Image Forest Transform (IFT) and has been successfully used to solve many classification problems, like weather prediction, brain imaging, Content-Based Image Retrieval (CBIR) and others.

The OPF-based classifiers have the following advantages: naturally multi-class; fast fitting and predicting; good accuracy; few parameters; suitable for supervised and unsupervised classification; and allows some superposition among clusters.

The goal of this project is to explore all potential of the OPF classifier, identifying its strengths and weaknesses, mapping its behavior in presence of noise and testing alternative ways of trainning it and of choosing its prototipes.

Annotation Tools

Recently modified pages
More