Scale-Space and Edge Detection Using Anisotropic Diffusion
IEEE Transactions on Pattern Analysis and Machine Intelligence
Identification of Growth Seeds in the Neonate Brain through Surfacic Helmholtz Decomposition
IPMI '09 Proceedings of the 21st International Conference on Information Processing in Medical Imaging
Variational method for super-resolution optical flow
Signal Processing
Feature detection and tracking in optical flow on non-flat manifolds
Pattern Recognition Letters
| Hi-index | 0.00 |
The critical points (also known as phase singularities) in the heart reflect the pathological change of the heart tissue, and hence can be used to describe and analyze the dynamics of the cardiac electrical activity. As a result, the detection of these critical points can lead to correct understanding and effective therapy of the tachycardia. In this paper, we propose a novel approach to address this problem. The proposed approach includes four stages: image smoothing, motion estimation, motion decomposition, and detection of the critical points. In the image smoothing stage, the noisy cardiac optical data are smoothed using anisotropic diffusion equation. The conduction velocity fields of the cardiac electrical patterns can then be estimated from two consecutive smoothed images. Using the recently developed discrete Hodge-Helmholtz motion decomposition technique, the curl-free and divergence-free potential surfaces of an estimated velocity field are extracted. Finally, hierarchically searching the minima and maxima on the potential surfaces, the sources, sinks, and rotational centers are located with high accuracy. Experimental results with four real cardiac videos show that the proposed approach performs satisfactorily, especially for the cardiac electrical patterns with simple propagations.