Spelling suggestions: "subject:"280406 amathematical 8oftware"" "subject:"280406 amathematical 5oftware""
1 |
Some extensions to support vector machinesMasters, A. L. Unknown Date (has links)
No description available.
|
2 |
Cognition driven deformation modellingJanke, A. Unknown Date (has links)
No description available.
|
3 |
Cognition driven deformation modellingJanke, A. Unknown Date (has links)
No description available.
|
4 |
Cognition driven deformation modellingJanke, Andrew Lindsay Unknown Date (has links)
This thesis describes the development of a model of cerebral atrophic change associated with neurodegeneration. Neurodegenerative diseases such as Alzheimer's dementia present a significant health problem within the elderly population. Effective treatment relies upon the early detection of anatomic change, and the subsequent differential diagnosis of the disorder from other closely related neurological conditions. Importantly, this also includes the investigation of the relationship between atrophic change and cognitive function. In unison with the growth in neuroimaging technology, myriad methodologies have been developed since the first quantitative measures of atrophic change were deduced via manual tracing. Subsequently, automated region of interest analysis, segmentation, voxel-based morphometry and non-linear registration have all been used to investigate atrophy. These methods commonly report findings of ventricular enlargement and temporal lobe change in AD and other dementias. Whilst these results are accurate indicators of atrophy, they are largely non-specific in their diagnostic utility. In addition, the aforementioned methods have been employed to discern change observed at discrete intervals during a disease process. In order to gain a greater understanding of the temporal characteristics of changes that occur as a result of atrophy, a deformation modelling method that allows the continuous tracking of these changes in a cohort of AD patients and elderly control subjects is presented in this thesis. Deformation modelling involves non-linear registration of images to investigate the change that is apparent between two or more images. The non- linear registration results are analysed and presented via three metrics: local volume loss (atrophy); volume (CSF) increase; and translation (interpreted as representing collapse of cortical structures). Changes observed in the analyses in this thesis are consistent with results from neuro-anatomical studies of AD. Results using the more traditional methods of analysis are presented for comparative purposes.
|
5 |
Cognition driven deformation modellingJanke, Andrew Lindsay Unknown Date (has links)
This thesis describes the development of a model of cerebral atrophic change associated with neurodegeneration. Neurodegenerative diseases such as Alzheimer's dementia present a significant health problem within the elderly population. Effective treatment relies upon the early detection of anatomic change, and the subsequent differential diagnosis of the disorder from other closely related neurological conditions. Importantly, this also includes the investigation of the relationship between atrophic change and cognitive function. In unison with the growth in neuroimaging technology, myriad methodologies have been developed since the first quantitative measures of atrophic change were deduced via manual tracing. Subsequently, automated region of interest analysis, segmentation, voxel-based morphometry and non-linear registration have all been used to investigate atrophy. These methods commonly report findings of ventricular enlargement and temporal lobe change in AD and other dementias. Whilst these results are accurate indicators of atrophy, they are largely non-specific in their diagnostic utility. In addition, the aforementioned methods have been employed to discern change observed at discrete intervals during a disease process. In order to gain a greater understanding of the temporal characteristics of changes that occur as a result of atrophy, a deformation modelling method that allows the continuous tracking of these changes in a cohort of AD patients and elderly control subjects is presented in this thesis. Deformation modelling involves non-linear registration of images to investigate the change that is apparent between two or more images. The non- linear registration results are analysed and presented via three metrics: local volume loss (atrophy); volume (CSF) increase; and translation (interpreted as representing collapse of cortical structures). Changes observed in the analyses in this thesis are consistent with results from neuro-anatomical studies of AD. Results using the more traditional methods of analysis are presented for comparative purposes.
|
Page generated in 0.0666 seconds