Streamlined Quantitative Evaluation of Brain Structural Anomalies in MRI Images

Zeljkovic V.; Quispe N.Y.A.; Druzgalski C.; Mayorga P. · Pan American Health Care Exchanges, PAHCE (2022) · DOI: 10.1109/gmepe/pahce55115.2022.9757792

Resumen

There are different conditions which may affect structure of the brain including aneurysm, brain tumor or brain injury, multiple sclerosis, and other. Evaluation of the brain as a soft tissue is very suitable to MRI as these scans provide more details, although CT is also often used as a quick tool for an initial diagnosis. In particular, the tumor tissue mainly appears as brighter shades than the rest of the regions in the brain. Detection of a brain tumor is a tedious task and its identification often involves segmentation of brain images, extraction of brain features and classification of abnormality in the MRI image. Based on this reflection we propose algorithm for brain structural changes documentation of selected image(s) and a tool assisting doctors in accelerated diagnosis and quantification of the degree of abnormality. Specifically, presented approach includes at the initial stage feature enhancement including minimizing gray-scale shade variations. In its final stage this allows color coding of presumptive tumor or other regions of interest (ROI) areas and gives an opportunity of comparative evaluation of tumor growth or reduction of its size. This is possible by a graphical and accompanied numerical presentation in terms of the percentage of the area classified as a tumor or other anomaly in reference to the total area of the brain reflected on a selected MRI image. Although this paper focuses on brain tumor, developed technique can be applied to images documenting other structural brain anomalies.