Showing papers by "Qi Zhang published in 2007"
23 May 2007
TL;DR: The results show that this scheme is superior to traditional GVF snake in terms of boundary localization, and the robustness of the algorithm is enhanced, since the GVf snake is combined with discrete wavelet transform to deform the contour in multiscale images.
Abstract: The extraction of luminal borders (contours) from intravascular ultrasound (IVUS) images is helpful for the diagnosis of coronary artery diseases. A novel scheme for contour extraction is proposed in this paper, based on GVF snakes and wavelet transform. To solve the two difficulties of the traditional GVF snake, i.e. the contour initialization and the suppression of noise and artifact interference, there are two improvements in our proposed scheme. First, the procedure is made into full automation, with adopting the characteristics of the image sequences to produce an initial contour. Secondly, the robustness of the algorithm is enhanced, since the GVF snake is combined with discrete wavelet transform to deform the contour in multiscale images. The proposed scheme is verified on both the synthetic images and real IVUS images. The results show that this scheme is superior to traditional GVF snake in terms of boundary localization.
1 citations
03 Jun 2007
TL;DR: A novel scheme based on the generalized relevance learning vector quantization (GRLVQ) using multiple parameters ( features) to gradually prune the unimportant features according to their weighting factors and proved to be superior to the traditional single-parameter method.
Abstract: There are fewer effective methods to accurately discriminate the coronary microcirculatory dysfunction from the normal coronary microcirculation Rather than traditional approaches only considering a single hemodynamic parameter, a novel scheme is proposed based on the generalized relevance learning vector quantization (GRLVQ) using multiple parameters (features) Naturally integrating the tasks of feature selection and classification, this scheme circularly adopts GRLVQ to gradually prune the unimportant features according to their weighting factors In each circulation, the prototypes are generated for classification and the classification accuracy is obtained Finally, the feature subset with the highest classification accuracy is selected and the corresponding classifier is also achieved This approach not only simplifies the classifier but also enhances the classification performance The method is verified on the physiological data collected from animals, and proved to be superior to the traditional single-parameter method