Symmetry Detection in Brain Image Analysis

Symmetry Detection in Brain Image Analysis

Surani Anuradha Jayasuriya, Alan Wee-Chung Liew, Phillip Sheridan
DOI: 10.4018/978-1-4666-5888-2.ch554
(Individual Chapters)
No Current Special Offers

Chapter Preview



Role of Computer-Aided Image Analysis Techniques in Neurological Practice

Neurological diseases are hard to diagnose due to their complex nature. Studies have reported that severe cortical volume loss (atrophy) is a consistent neuro-pathological finding in many brain disorders (Mouton et al., 1998). Absolute volumes and the spatial distribution of brain atrophy are of major interest when studying the course of neurodegenerative diseases, such as Multiple Sclerosis, Alzheimer’s disease (AD), and Schizophrenia.

The imaging modalities most often used for diagnosis of brain diseases are magnetic resonance imaging (MRI) and X-ray computed tomography (CT). While neurological studies have greatly benefitted by high resolution in vivo neuroimaging, assessment of these images are generally performed manually by radiologists. Manual processing of a large number of scans is a time consuming and tedious task. Moreover, the image-derived measurements taken by manual delineation of structures of interest by a radiologist can be affected by inter- and intra-operator variability and poor reproducibility. Various artifacts in different imaging modalities can also compromise accurate image interpretation. It becomes important to develop objective methods for accurate measurement of brain atrophy or localization of pathology.

Computers excel in performing quantitative tasks. In neuroimaging, one of the questions that computers can answer more reliably than humans is in regards to the precise quantification of the degree to which a tissue type has changed (Hahn, 2010). In order to facilitate accurate detection of abnormalities and quantification, intensive research has been conducted on automatic medical image analysis. Advantages of an automated method include objectivity, reproducibility and the capability to process a large number of images (Jayasuriya and Liew, 2012).

However, despite several decades of research, automatic analysis of brain images still remains a challenging task due to the anatomical complexity of the brain as well as the various noise artifacts in brain images. For example, in magnetic resonance images, artifacts such as intensity non-uniformity (Liew and Yan, 2006) give rise to smooth intensity variation across the image (Figure 1). Recently, attention has turned towards incorporating knowledge gained from brain anatomy and principles used by medical experts. One strategy is to utilize brain symmetry which is routinely employed by expert clinicians in order to identify brain pathologies, such as stroke, tumor and atrophy. This approach compares suspicious tissues to healthy tissues in other regions of the brain.

Figure 1.

An image of a human brain: It is highly symmetrical. The deep groove, known as the inter-hemispheric fissure separates the two hemispheres.


Key Terms in this Chapter

Mid-Sagittal Plane (MSP): The plane that separates the left and right brain hemispheres which also lines up with the sagittal plane of the human body.

Computer-Aided Medical Image Analysis: Procedures in medicine in which computer tools assist doctors in the interpretation of medical images.

Brain Symmetry: Neuroanatomical similarity between the two brain hemispheres.

Prior Knowledge: The knowledge that stems from previous experience.

Neuroimaging: The use of imaging techniques to visualize the structure or function of the brain.

Inter-Hemispheric Fissure: The deep groove that separates the two brain hemispheres. Also known as longitudinal fissure.

Axial Slices: A plane section made by cutting the brain at right angles to the MSP.

Pathology Detection: Identification of any abnormality in the brain.

Complete Chapter List

Search this Book: