Research project utilises computer vision, AI techniques to assess brain abnormalities
The proposed method determined the status of the brain by comparing the two lobes of the brain with each other and decides the presence of abnormalities in it based on the results of the comparison
With the development of technology, there has been an array of research to enhance all fields of life.
‘Utilising Computer Vision and AI in Diagnostic Brain Abnormalities’ by Dr Mohammad A al Azawi, Associate Professor of Artificial Intelligence and Computer Vision at Al Buraimi University College, is one of the research projects that contributed to improving the medical field.
In this research project, Dr Mohammad stated that the study was part of a project that employs advanced computer vision and AI techniques for the analysis and processing of medical imaging, mentioning that the work delved into the domain of medical image processing, particularly focusing on magnetic resonance image (MRI) processing within the realm of computer-aided diagnostic (CAD) systems.
Dr Mohammad emphasised that the application of computer vision and AI in analysing medical images and assessing brain anomalies is a crucial aspect. However, the inherent variability in brain shapes among individuals poses a challenge for computer-based diagnoses.
To address this challenge, Dr Mohammad introduced a novel method leveraging the symmetry of the brain’s two lobes through this study. The study utilised the fact that the two lobes of the brain are symmetric to decide if there are any abnormalities as tumours cause a deformation in the shape of one of the lobes, which affects this symmetry.
The proposed method determined the status of the brain by comparing the two lobes of the brain with each other and decides the presence of abnormalities in it based on the results of the comparison.
According to Dr Mohammad, the comparison process involved analysing various features extracted from images, including colour and texture. The proposed algorithm was applied to 300 images from standard datasets, yielding highly satisfactory results where the precision, recall, and accuracy reached 95.3 per cent, 94.7 per cent, and 95 per cent respectively. The study, which was published in the journal Multimedia Tools and Applications, suggested that leveraging the inherent symmetry of the brain’s lobes was an effective approach for computer-based diagnosis of abnormalities.