Title :
MVN_CNN and FCNN for endocardial edge detection
Author :
Ketout, Hussin ; Gu, Jason ; Horne, Gabrielle
Author_Institution :
Electr. & Comput. Eng., Dalhousie Univ., Halifax, NS, Canada
Abstract :
In this paper, Fuzzy Cellular Neural Networks (FCNN) endocardial edge detection is proposed. The echocardiographic image is preprocessed to enhance the contrast and smoothness by utilizing MVN_CNN filtering. FCNN is applied to the smoothed image to extract the heart boundaries. Fuzzy min and max functions are employed. The comparison was made between Fuzzy, CNN and FCNN edge detectors. The FCNN approach showed better results for extracting the LV endocardial edges. Some experimental results are given for different echocardiographic images.
Keywords :
echocardiography; edge detection; fuzzy neural nets; medical image processing; FCNN filtering; Fuzzy Cellular Neural Network; MVN_CNN filtering; echocardiographic image; endocardial edge detection; heart boundary; Cellular neural networks; Equations; Image edge detection; Mathematical model; Maximum likelihood detection; Nonlinear filters; CNN; Echocardiography; Endocardial; FCNN; Fuzzy; MVN_CNN; edge detection;
Conference_Titel :
Biomedical Engineering (MECBME), 2011 1st Middle East Conference on
Conference_Location :
Sharjah
Print_ISBN :
978-1-4244-6998-7
DOI :
10.1109/MECBME.2011.5752102
