1Knowledge Systems Knowledge Systems Laboratory Laboratory Advances in the Use of Advances in the Use of Neurophysiologycally-based Neurophysiologycally-based Fusion for Visualization and Fusion for Visualization and Pattern Recognition of Medical Pattern Recognition of Medical Imagery Imagery M. Aguilar, J. R. New and E. M. Aguilar, J. R. New and E. Hasanbelliu Hasanbelliu Knowledge Systems Laboratory Knowledge Systems Laboratory MCIS Department MCIS Department Jacksonville State University Jacksonville State University Jacksonville, AL 36265 Jacksonville, AL 36265
23
Embed
M. Aguilar, J. R. New and E. Hasanbelliu Knowledge Systems Laboratory MCIS Department
Advances in the Use of Neurophysiologycally-based Fusion for Visualization and Pattern Recognition of Medical Imagery. M. Aguilar, J. R. New and E. Hasanbelliu Knowledge Systems Laboratory MCIS Department Jacksonville State University Jacksonville, AL 36265. Outline. Introduce Med-LIFE. - PowerPoint PPT Presentation
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
11Knowledge Systems Knowledge Systems
LaboratoryLaboratory
Advances in the Use of Advances in the Use of Neurophysiologycally-based Neurophysiologycally-based Fusion for Visualization and Fusion for Visualization and
Pattern Recognition of Medical Pattern Recognition of Medical ImageryImagery
M. Aguilar, J. R. New and E. M. Aguilar, J. R. New and E. HasanbelliuHasanbelliu
Knowledge Systems LaboratoryKnowledge Systems LaboratoryMCIS DepartmentMCIS Department
Jacksonville State UniversityJacksonville State UniversityJacksonville, AL 36265Jacksonville, AL 36265
22Knowledge Systems Knowledge Systems
LaboratoryLaboratory
OutlineOutline Introduce Med-LIFE.Introduce Med-LIFE. Revisit 3D image fusion architecture.Revisit 3D image fusion architecture. Compare 2D and 3D fusion results.Compare 2D and 3D fusion results. Fusion for segmentation and pattern Fusion for segmentation and pattern
Supervised learning where training data Supervised learning where training data is selected by user/expert (Waxman et is selected by user/expert (Waxman et al).al).
Results assessed and corrected by user.Results assessed and corrected by user. Fuzzy ARTMAP neural network for fast Fuzzy ARTMAP neural network for fast
and stable learning.and stable learning. Address order sensitivity by introducing Address order sensitivity by introducing
N voters trained with alternate ordering N voters trained with alternate ordering of the training data.of the training data.
Heterogeneous VotingHeterogeneous Voting Train 3 Fuzzy ARTMAP systems with Train 3 Fuzzy ARTMAP systems with
parameters as before (different data parameters as before (different data orderings)orderings)
Train remaining 2 systems with all Train remaining 2 systems with all parameters as in the 3parameters as in the 3rdrd system system except for except for VigilanceVigilance (which is a (which is a threshold measure that controls the threshold measure that controls the sensitivity of the system).sensitivity of the system).
1717Knowledge Systems Knowledge Systems
LaboratoryLaboratory
Homogeneous vs. Homogeneous vs. Heterogeneous VotersHeterogeneous Voters
5 Homogeneous Voters 5 Heterogeneous Voters
1818Knowledge Systems Knowledge Systems
LaboratoryLaboratory
2D vs. 3D Fusion 2D vs. 3D Fusion Segmentation ResultsSegmentation Results
of 2D and 3D fusion.of 2D and 3D fusion. Preliminary learning segmentation results Preliminary learning segmentation results
indicate robustness across slices and cases.indicate robustness across slices and cases. Demonstrated superior performance of 3D Demonstrated superior performance of 3D
fusion for both visualization and pattern fusion for both visualization and pattern recognition.recognition.