Loading...
Please wait, while we are loading the content...
Similar Documents
Comparison of three gaussian mixture modeling and spatial encoding methods for segmenting human brain mri.
| Content Provider | CiteSeerX |
|---|---|
| Author | Zeydabadi-Nejad, Mahmood Zoroofi, Reza A. Soltanian-Zadeh, Hamid |
| Abstract | Abstract:- We propose a method to improve performance of image segmentation methods that are based on Gaussian mixture models and spatial encoding, and then compare them on T1-weighted MR images. Three approaches are considered for this study. The first one considers a simple neighborhood system to encode the spatial relation between image pixels. The second one uses a multi-resolution neighborhood system to take into account the spatial information. The third one is a new segmentation algorithm we proposed and uses a wavelet based multi-resolution scheme to include the spatial consideration into the standard Gaussian mixture model. The methods are tested on nearly 4000 synthetic and real images. |
| File Format | |
| Access Restriction | Open |
| Subject Keyword | Gaussian Mixture Modeling Spatial Encoding Method Human Brain Mri Multi-resolution Scheme Multi-resolution Neighborhood System Gaussian Mixture Model Real Image Spatial Information Image Segmentation Method Standard Gaussian Mixture Model Simple Neighborhood System New Segmentation Spatial Consideration Image Pixel T1-weighted Mr Image Spatial Encoding Spatial Relation |
| Content Type | Text |
