Researcher from CEFOP development algorithm to rapid analysis of brain imaging

Neuronal interaction and early detection of brain diseases inspired the study called “Fast segmentation of cerebral white matter fibers”. The research, realized by student Nicole Labra, from Center for Optics and Photonics (CEFOP), allowed her to reach the Master of Science in Electric Engineering degree at Universidad de Concepción.

The research, lead by Professor Dr. Miguel Figueroa, has sought to create a new algorithm for rapid classification of white matter fibers for interactive applications, based it in an algorithm previously developed by professor Dr. Pamela Guevara. The new algorithm will allow a better study of interaction behavior of present neurons at the moment of a pulse transmitted between them.

IMG_2550The young researcher explained the reasons to focus the study. “One of my tutors, Pamela Guevara, worked this issue in France together with Professor (Miguel) Figueroa. They were seeking to optimize their algorithm, because the process was rather slow. This did not allowed to use it in any hospital, then it must be necessary and susceptible to joining interactive applications, for anyone interested in the study of the brain can use it “, said.

One of the most relevant points of the study conclusions, highlighted by the evaluation committee chaired by Dr. Sergio Torres, is that through the algorithm, it is possible get to same results as with Dr. Guevara’s, but in a more efficient way. This is possible because of the new algorithm and the exploitation of parallel execution capabilities, available in contemporary microprocessors and graphics cards. In that regard, Nicole highlighted that research has important practical applications. “Applications of these results are given with sick patients, analyzing databases, which are very large. Therefore, it was necessary a faster algorithm. The challenge now is to incorporate applications or create new applications” said biomedical civil engineer, who is currently applying for a job in France, since its objective is further developed in the research area of this field.

For his part, Professor Dr. Miguel Figueroa, highlighted the results of research “allow extracting useful information from large datasets of brain structure. The algorithm developed by Nicole allows us to know information about anatomical human brain, which is extremely useful for specialists who perform the analysis. Furthermore, this algorithm can obtain this information almost instantaneously, varying analysis parameters interactively. This is a great challenge, because the data sets we work with are large, so that traditional algorithms are too slow. As an illustration, the algorithm developed by Nicole reduces analysis time of one million fibers from almost two hours of the original formulation of the method, to less than one second. These results are the product of a high-quality work developed by Nicole, which showed a remarkable initiative, creativity, and ability to solve problems”, emphasized professor Figueroa from CEFOP.



Categories: News