Developed software that allows you to build a three-dimensional image of the brain from magnetic resonance imaging (MRI) images.
Alexander Grishchenko, a fourth—year student at the Faculty of Automation and Computer Engineering at NSTU-NETI, and Anton Gerasimov, a graduate student at the Department of Data Acquisition and Processing Systems, have developed a program that allows you to create a three-dimensional image of the brain. Previously, it was possible to consider only areas on the surface of the cortex.
First of all, the software is intended for clinical practice as an auxiliary tool for visualizing MRI data in diagnosis, surgery planning, and monitoring the dynamics of pathologies. The software can also serve as a platform for creating more specialized analytical tools.
To build a three-dimensional image of the brain, a series of MRI scans is required, which together represent a three-dimensional matrix of three—dimensional voxel pixels. They display the intensity of reactions of hydrogen atoms in the body to a magnetic field, from which noise is removed — voxels with very low and very high intensities. As a result, according to Alexander Grishchenko, "clean" head scan data is obtained.
To build an image of the brain, it is necessary to isolate it from the received data array. According to the developer, the brain extraction algorithm is based on the method of a deformable surface model, when a tessellated sphere (a sphere made of a huge number of polygons) is placed in the center of the head on MRI images.
"Near each vertex of the polygons that make up the sphere, a comparison of intensities takes place, on the basis of which a decision is made on whether the brain model will "expand" or "shrink" at this point. The algorithm constantly adjusts the position of the vertices so that they are evenly spaced, without wrinkles or breaks. This allows you to maintain a realistic and holistic shape of the brain throughout the entire deformation process. These steps are repeated many times to achieve maximum model accuracy. In this way, the contour of the brain is formed, which already determines its volume," said Alexander Grishchenko.
"The program allows you not only to isolate the brain and build its three—dimensional image, but also, if necessary, to identify individual areas of the brain to analyze its internal structures — areas of interest," said the developer and added that in the future he plans to calculate the mass of the brain from MRI images based on software already running.
Earlier, NSTU-NETI developed a voice screening system to assess the neurological status of patients with Parkinson's disease. The unique monitoring method has no analogues in Russia and is already being used at the Federal Center for Neurosurgery in Novosibirsk.