PicSOM—content-based image retrieval with self-organizing maps
Pattern Recognition Letters - Selected papers from the 11th scandinavian conference on image analysis
Handbook of medical imaging
Principles of data mining
Self-Organizing Maps
Face Recognition Using Laplacianfaces
IEEE Transactions on Pattern Analysis and Machine Intelligence
Hi-index | 0.00 |
This paper addresses the issue of a content-based information retrieval system that works on fMRI images from neuroscientific journal publications. We present a general framework for automatic extraction, characterisation and classification of fMRI images, based on their functional properties. The proposed method identifies the section of each of those images, by morphological processing, and estimates the coordinates of the brain activated regions, in relation to a standard reference template using locality preserving projections. Those regions are then segmented, and their physical and geometrical properties evaluated. We formulate a feature vector based on these characteristics, and cluster the images and corresponding journal publications using self organizing maps.