Comparison Of Brain Tissue Classification Algorithms With The Use Of Multimodal Efficiency Measure
|Original title:||Comparison Of Brain Tissue Classification Algorithms With The Use Of Multimodal Efficiency Measure|
|Authors:||Eva Janoušová, Daniel Schwarz|
|Field:||Use of computers, robotics and its application|
|Type:||Article in Proceedings|
|Keywords:||computational neuroanatomy;image analysis;classification;segmentation|
A multimodal measure for comparison of efficiency of classification algorithms is proposed. The algorithms selected for brain tissue classification in 3-D MRI images are the well-known k-NN and k-means classifiers which were adapted to the image data in the common stereotaxic space. The efficiency measure combines Jaccard coefficient, modified Rand index and Connectivity coefficient. Results are presented in 3-D plots with the use of ellipsoids instead of frequently used 1-D plots or tables.