TY  - THES
AU  - Aura Hernandez-Sabate
PY  - 2005//
TI  - Automatic adventitia segmentation in IntraVascular UltraSound images
CY  - 08193 Bellaterra, Barcelona (Spain)
N2  - A usual tool in cardiac disease diagnosis is vessel plaque assessment by analysis of IVUS sequences. Manual detection of lumen-intima, intima-media and media-adventitia vessel borders is the main activity of physicians in the process of plaque quantification. Large variety in vessel border descriptors, as well as, shades, artifacts and blurred response due to ultrasound physical properties troubles automated media-adventitia segmentation. This experimental work presents a solution to such a complex problem. The process blends advanced anisotropic filtering operators and statistic classification techniques, achieving an efficient vessel border modelling strategy. First of all, we introduce the theoretic base of the method. After that, we show the steps of the algorithm, validating the method with statistics that show that the media-adventitia border detection achieves an accuracy in the range of inter-observer variability regardless of plaque nature, vessel geometry and incomplete vessel borders. Finally, we present a little Matlab application to the automatic media-adventitia border.
L1  - http://refbase.cvc.uab.es/files/Her2005.pdf
N1  - IAM;
ID  - Aura Hernandez-Sabate2005
U1  - Masters thesis
ER  -