Title :
Auto detection of brain ventricles using Hausdorff distance
Author :
Sim, K.S. ; Ong, M.K. ; Chong, S.S. ; Ng, J.T. ; Tso, C.P. ; Choo, S.L. ; Rozalina, A.H.
Author_Institution :
Fac. of Eng. & Technol., Multimedia Univ., Ayer Keroh, Malaysia
fDate :
Nov. 30 2010-Dec. 2 2010
Abstract :
Brain plays an important role in human anatomy. The brain ventricular system can often be affected by different kinds of brain lesion, to the extent of creating an imbalance problem in the brain system. Hence, it is useful to develop a method to check the brain condition and to detect the existence of ventricles as well. A template of the ventricle is first created. Then ventricle detection in a slice of brain scan is done by using the Hausdorff distance. The presence of ventricles is smoothed by a median filter. Then, the ventricles are split into left and right halves through the computed centroid. Ratio is calculated to detect the brain abnormality. Result showed that 50% of the patients were detected accurately based on the ventricles.
Keywords :
brain; image matching; medical image processing; neurophysiology; Hausdorff distance; brain abnormality detection; brain lesions; brain scan slice; brain ventricle autodetection; brain ventricular system; Accuracy; Brain models; Pixel; Brain ventricles; Hausdorff distance; brain lesion; template matching;
Conference_Titel :
Biomedical Engineering and Sciences (IECBES), 2010 IEEE EMBS Conference on
Conference_Location :
Kuala Lumpur
Print_ISBN :
978-1-4244-7599-5
DOI :
10.1109/IECBES.2010.5742228