Biomedical subjects

Jussi Tohka

Publications and source records attributed to Jussi Tohka.

5 recordsLinked to original sources

PubMed2005

Automatic extraction of brain surface and mid-sagittal plane from PET images applying deformable models.

In this study, we propose and evaluate new methods for automatic extraction of the brain surface and the mid-sagittal plane from functional positron emission tomography (PET) images. Designing methods for these segmentation tasks is challenging because the spatial distribution of intensity values in a PET image depends on the applied radiopharmaceutical and the contrast to noise ratio in a PET image is typically low. We extracted the brain surface with a deformable model which is based on a global optimization algorithm. The global optimization allows reliable automation of the extraction task. Based on the extracted brain surface, the mid-sagittal plane was determined. The method was tested with the image of the Hoffman brain phantom (FDG) and the images from the brain studies with the FDG (17 images) and the C11-Raclopride tracers (4 images). In addition to the brain surfaces, we applied the deformable model for extraction of the coarse cortical structure based on the tracer uptake from FDG-PET brain images. The proposed segmentation methods provide a promising direction for automatic processing and analysis of PET brain images.

Adult↗

PubMed2005

Object-oriented implementation for the dual surface minimisation algorithm.

The automatic surface extraction from volumetric images is important for the medical image analysis. The delineation task is challenging especially with functional imaging where images are noisy and the content of image depends on the study. We have earlier developed a new dual surface minimisation (DSM) algorithm for the optimisation of deformable surfaces and applied it successfully to fully automatic surface extraction from positron emission tomography images. We present a fast object-oriented implementation for the DSM algorithm with a conventional command line interface. The DSM software is a freely available tool for the Internet community at http://www.cs.uta.fi/research/software/. The implementation of the DSM algorithm provides new opportunities to extend experiments, to develop and to study new applications with noisy volumetric images emerging from medical or other applications.

Algorithms↗

PubMed2004

Fast and robust parameter estimation for statistical partial volume models in brain MRI.

Due to the finite spatial resolution of imaging devices, a single voxel in a medical image may be composed of mixture of tissue types, an effect known as partial volume effect (PVE). Partial volume estimation, that is, the estimation of the amount of each tissue type within each voxel, has received considerable interest in recent years. Much of this work has been focused on the mixel model, a statistical model of PVE. We propose a novel trimmed minimum covariance determinant (TMCD) method for the estimation of the parameters of the mixel PVE model. In this method, each voxel is first labeled according to the most dominant tissue type. Voxels that are prone to PVE are removed from this labeled set, following which robust location estimators with high breakdown points are used to estimate the mean and the covariance of each tissue class. Comparisons between different methods for parameter estimation based on classified images as well as expectation--maximization-like (EM-like) procedure for simultaneous parameter and partial volume estimation are reported. The robust estimators based on a pruned classification as presented here are shown to perform well even if the initial classification is of poor quality. The results obtained are comparable to those obtained using the EM-like procedure, but require considerably less computation time. Segmentation results of real data based on partial volume estimation are also reported. In addition to considering the parameter estimation problem, we discuss differences between different approximations to the complete mixel model. In summary, the proposed TMCD method allows for the accurate, robust, and efficient estimation of partial volume model parameters, which is crucial to a variety of brain MRI data analysis procedures such as the accurate estimation of tissue volumes and the accurate delineation of the cortical surface.

Artifacts↗

PubMed2003

Delineation of brain structures from positron emission tomography images with deformable models.

Segmentation of positron emission tomography (PET) images is a difficult task. In this study, we propose a new method for delineation of brain structures according to the tracer uptake. The method is based on a new deformable model which is particularly designed for extracting surfaces automatically from noisy images. The automation is achieved by using a global optimization algorithm for minimizing the energy of the deformable model. As an example, the coarse cortical structure was extracted from FDG PET brain images by delineating first the brain surface and then the white matter surface. We have tested the method with the image of the brain phantom and images from a small number (N = 17) of FDG brain studies. The cortical structure was automatically and reliably found from all the images. The proposed method provides new opportunities for automatic and repeatable structure extraction applicable for regional quantification of the tracer uptake.

Brain↗

PubMedDate not supplied

Automated diagnosis of brain tumours astrocytomas using probabilistic neural network clustering and support vector machines.

A computer-aided diagnosis system was developed for assisting brain astrocytomas malignancy grading. Microscopy images from 140 astrocytic biopsies were digitized and cell nuclei were automatically segmented using a Probabilistic Neural Network pixel-based clustering algorithm. A decision tree classification scheme was constructed to discriminate low, intermediate and high-grade tumours by analyzing nuclear features extracted from segmented nuclei with a Support Vector Machine classifier. Nuclei were segmented with an average accuracy of 86.5%. Low, intermediate, and high-grade tumours were identified with 95%, 88.3%, and 91% accuracies respectively. The proposed algorithm could be used as a second opinion tool for the histopathologists.

Algorithms↗