The overall goal of this research is to develop methods and algorithms to investigate the severity of Traumatic brain injury (TBI) and to estimate the intracranial pressure (ICP) level non-invasively. Brain x-ray computed tomography (CT) images and artificial intelligence methods are employed to estimate the level of ICP. Fully anisotropic complex wavelet transform features are proposed to extract directional textural features from brain images. Different feature selection and classification methods are tested to find the optimal feature vector and estimate the ICP using support vector regression. By using systematic feature extraction, selection and classification, promising results on ICP estimation are achieved. The results also indicate the reliability of the proposed algorithm. In the following, case-based finite element (FE) models are extracted from CT images using Matlab, Solidworks, and Ansys software tools. The ICP estimation obtained from image analysis is used as an input to the FE modeling to obtain stress/strain distribution over the tissue. Three in-plane modeling approaches are proposed to investigate the effect of ICP elevation on brain tissue stress/strain distribution. Moreover, the effect of intracranial bleeding on ICP elevation is studied in 2-D modeling. A mathematical relationship between the intracranial pressure and the maximum strain/stress over the brain tissue is obtained using linear regression method. In the following, a 3-D model is constructed using 3 slices of brain CT images. The effect of increased ICP on the tissue deformation is studied. The results show the proposed framework can accurately simulate the injury and provides an accurate ICP estimation non-invasively. The results from this study may be used as a base for developing a non-invasive procedure for evaluating ICP using FE methods.
| Identifer | oai:union.ndltd.org:vcu.edu/oai:scholarscompass.vcu.edu:etd-3659 |
| Date | 10 February 2012 |
| Creators | Seyed, Aghazadeh Babak |
| Publisher | VCU Scholars Compass |
| Source Sets | Virginia Commonwealth University |
| Detected Language | English |
| Type | text |
| Format | application/pdf |
| Source | Theses and Dissertations |
| Rights | © The Author |
Page generated in 0.0021 seconds