Learn More
A novel color image segmentation method based on finite Gaussian mixture model is proposed in this paper. First, we use EM algorithm to estimate the distribution of input image data and the number of mixture components is automatically determined by MML criterion. Then the seg-mentation is carried out by clustering each pixel into appropriate component(More)
Linear representations are widely used to reduce dimension in applications involving high dimensional data. While specialized procedures exist for certain optimality criteria, such as principle component analysis (PCA) and Fisher discrim-inant analysis (FDA), they can not be generalized for more general criteria. To overcome this fundamental limitation,(More)
Cytoplasmic dynein 1 is fundamentally important for transporting a variety of essential cargoes along microtubules within eukaryotic cells. However, in mammals, few mutants are available for studying the effects of defects in dynein-controlled processes in the context of the whole organism. Here, we deleted mouse Dlic1 gene encoding DLIC1, a subunit of the(More)
— A novel image texture classification method based on finite Gaussian mixture models of sub-band coefficients is proposed in this paper. In the method, an image texture is first decomposed into several sub-bands, then the marginal density distribution of coefficients in each sub-band is approximated by Gaussian mixtures. The Gaussian component parameters(More)
To develop a computer-aided diagnostic scheme of the CT in the diagnosis of lung cancer based on artificial neural networks (ANN) to assist radiologists in distinguishing malignant from benign pulmonary nodules. 117 CT images of pulmonary nodules (58 benign and 59 malignant) were analyzed. 21 CT radiological features of each case were carefully selected and(More)
This paper is a contribution to the methodology of fully Bayesian inference in mul-tivariate Gaussian mixtures using the reversible jump Markov chain Monte Carlo algorithm. To make use of the spectral representation of symmetric positive definite matrix, we decompose covariance matrix into two parts: an eigenvector matrix and an eigenvalue matrix. We focus(More)
In this paper, a discriminative manifold learning method for face recognition is proposed which achieved the dis-criminative embedding the high dimensional face data into a low dimensional hidden manifold. Unlike the recently proposed LLE, Isomap and Eigenmap algorithms, which are based on reconstruction purpose, our method use the RCA algorithm to achieve(More)