Echo-cardiographic segmentation: Via feature-space clustering

Vidhyadhari Gondle, Jayanthi Sivaswamy
2011 2011 National Conference on Communications (NCC)  
Segmentation in echo-cardiographic images is a difficult task due to the presence of speckle noise, low contrast and blurring. We present a novel method based on clustering performed in the feature space. A noise robust image representation is obtained by computing a local feature descriptor at every pixel location. This descriptor is derived using the Radon-Transform to effectively characterise local image context. Next, an unsupervised clustering is performed in the feature space to segment
more » ... gions in the image. The performance of the proposed method is tested on both synthetic and real images. A comparison against well established feature descriptors is carried out to demonstrate the strengths and applicability of the proposed representation. Overall, the results indicate promise in the strategy of doing segmentation of noisy data in image descriptor space.
doi:10.1109/ncc.2011.5734776 fatcat:5mkvvh7ic5eyrk2rlb5p42s7ru