A multilevel approach to intelligent information filtering: model, system, and evaluation
ACM Transactions on Information Systems (TOIS)
Support Vector Machines for 3D Object Recognition
IEEE Transactions on Pattern Analysis and Machine Intelligence
Exploiting Image Indexing Techniques in DCT Domain
MINAR '98 Proceedings of the IAPR International Workshop on Multimedia Information Analysis and Retrieval
Unsupervised segmentation of medical images using DCT coefficients
VIP '05 Proceedings of the Pan-Sydney area workshop on Visual information processing
Classification for Breast MRI Using Support Vector Machine
CITWORKSHOPS '08 Proceedings of the 2008 IEEE 8th International Conference on Computer and Information Technology Workshops
A Support Vector Machine Based Algorithm for Magnetic Resonance Image Segmentation
ICNC '08 Proceedings of the 2008 Fourth International Conference on Natural Computation - Volume 03
Atlas guided identification of brain structures by combining 3d segmentation and SVM classification
MICCAI'06 Proceedings of the 9th international conference on Medical Image Computing and Computer-Assisted Intervention - Volume Part II
Multitraining Support Vector Machine for Image Retrieval
IEEE Transactions on Image Processing
MCBR-CDS'09 Proceedings of the First MICCAI international conference on Medical Content-Based Retrieval for Clinical Decision Support
Hi-index | 0.00 |
A project is described with the aim to develop a Computer-Aided Retrieval and Diagnosis of Alzheimer's disease. The domain of focus is Alzheimer's disease A manually curated MRI data set from the Alzheimer's Disease Neuroimaging Initiative (ADNI) project (http://www.loni.ucla.edu/ADNI/) was used for training and validation. The system's main function is to generate accurate matches for any given visual or textual query. The system gives an option to perform the matching based on a variety of feature-sets, extracted using an adaptation of a discrete cosine transform algorithm. Classification is conducted using Support Vector Machines. Finally, ranking of most accurate matches are generated by applying an Euclidean distance score. The overall system architecture follows a multi-level model, permitting performance analysis of components independently. Experimental results demonstrate that the system can produce effective results.