Brain volume estimation is a challenging area for research in medical domain. A quantitative analysis of the brain volume is very important for the diagnosis and tracking progression of various degenerative brain diseases in several clinical applications. This study aims to develop an automated method for estimating volume of the brain using MRI (Magnetic Resonance Imaging) data. In this paper, a novel approach of brain volume estimation using gradient based filters and volume refinement is proposed. The experiment is conducted on MRI volumetric data of total 20 patients. Live data of 10 patients is collected from Hospital and data of 10 patients is taken from publicly available Open Access Series of Imaging Studies (OASIS) dataset. The average accuracy of Brain volume estimation obtained through proposed approach with Gradient Magnitude Image Filter using OASIS data is 94.16% and using live data is 93.65%.Design of such approaches can assist radiologists by estimating brain volume and can help in determining existence and progression of neurodegenerative diseases.

Brain volume, Brain MRI analysis, Volume estimation