Presents an algorithm for image segmentation with irregular pyramids. Instead of starting with the original pixel grid, the authors first apply an adaptive Voronoi tessellation to the image. For irregular pyramid cons...
详细信息
Presents an algorithm for image segmentation with irregular pyramids. Instead of starting with the original pixel grid, the authors first apply an adaptive Voronoi tessellation to the image. For irregular pyramid construction the authors present a Hopfield neural network which controls the decimation process. The validity of the authors' approach is demonstrated by several examples in image segmentation.
This paper presents a new method for image compression by neural networks. First, we show that we can use neural networks in a pyramidal framework, yielding the so-called PCA pyramids. Then we present an image compres...
详细信息
This paper presents a new method for image compression by neural networks. First, we show that we can use neural networks in a pyramidal framework, yielding the so-called PCA pyramids. Then we present an image compres...
This paper presents a new method for image compression by neural networks. First, we show that we can use neural networks in a pyramidal framework, yielding the so-called PCA pyramids. Then we present an image compression method based on the PCA pyramid, which is similar to the Laplace pyramid and wavelet transform. Some experimental results with real images are reported. Finally, we present a method to combine the quantization step with the learning of the PCA pyramid.
This paper addresses the issue of tracking tubular objects, particularly blood vessels from MR images. A model-based approach is adopted. The generalized stochastic tube (GST) model is developed which is an extension ...
详细信息
This paper addresses the issue of tracking tubular objects, particularly blood vessels from MR images. A model-based approach is adopted. The generalized stochastic tube (GST) model is developed which is an extension of our previously proposed (1993) generalized tube (GT) model. Transitions among adjacent tubes are explicitly parameterized. Integrated with a bivariate Gaussian density function adopted to model the blood flow within cross sections, the GST model is applied to tracking blood vessels in MRA volumetric data. Experimental results on both synthetic data with different degrees of Gaussian noise and real MRA data demonstrated that simultaneously utilizing both models yields robust performance under noisy conditions.
A model-based approach is used for recognizing arterial blood vessels from MRA volumetric data. The modeling includes (1) a generalized stochastic tube model characterizing the structural properties of the vessels, an...
详细信息
A model-based approach is used for recognizing arterial blood vessels from MRA volumetric data. The modeling includes (1) a generalized stochastic tube model characterizing the structural properties of the vessels, and (2) a bivariate Gaussian function, modeling the expected cross sectional blood flow. This integrated model renders the recognition problem as a parameter estimation problem which is subsequently solved in a hierarchical fashion. Due to the descriptive representation for objects, a new visualization scheme for blood vessels is proposed that allows the observation of the blood flow of each cross section along a recognized vessel. Some experimental results from both synthetic data and real MRA volumes are given. The visualization of interior blood flow patterns for some vessels are also shown.< >
＜正＞Feature extraction is very important for the classifier design and the overall performance of *** recognition ***,due to the lack of theoretical guidances,feature extraction and classifier design are usually tre...
＜正＞Feature extraction is very important for the classifier design and the overall performance of *** recognition ***,due to the lack of theoretical guidances,feature extraction and classifier design are usually treated separately in current speech recognition *** *** proposes an approach to combine linear feature extraction with continuous density hidden Markov modeling（HMM） which is currently the most successful speech pattern classifier.A maximumlikelihood based algorithm is derived to iteratively train HMM parameters as well as the parameters of the feature *** algorithm is an exteusion of the Baum-Welcli parameter re-estimation algorithm for conventional HMMs and thus has a nice property of guara, nteed convergence.
In this paper we present a new shape normalization method that is invariant to shape translation, rotation and scaling. We define a visible area density function and an unvisible area density function for a planar sha...
详细信息
In this paper we present a new shape normalization method that is invariant to shape translation, rotation and scaling. We define a visible area density function and an unvisible area density function for a planar shape. Using these two functions we define a visible region center and an unvisible region center of the shape, respectively. When the visible and unvisible region centers of a shape locate at different positions they can be utilized as characteristic points to normalize the shape to a standard form. The normalizing process by use of the centers is presented. Experiments are executed on five groups of shapes with distortion of translation, rotation and scaling adding quantilization noise. The results show that the method is reasonable and available.
A neural network appraoch for classification using features extracted by a mapping is presented. When the number of sample dimensions is much larger than the number of classes and no deviations are given but the means...
详细信息
A neural network appraoch for classification using features extracted by a mapping is presented. When the number of sample dimensions is much larger than the number of classes and no deviations are given but the means of classes, a mapping from class space to a new one whose dimensions is exactly equal to the number of classes is proposed. The vectors in the new space are considered as the feature vectors to be inputted to a neural network for classification. The property that the mapping does not change the separability of the original classification problem is given. Simulation results for object recognition are presented.
A new update criterion, which is called eliminating highest error (EHE) criterion, is presented in the paper. The Hopfield neural network taking the threshold function as the neuron's output function is essentiall...
详细信息
A new update criterion, which is called eliminating highest error (EHE) criterion, is presented in the paper. The Hopfield neural network taking the threshold function as the neuron's output function is essentially unstable when it is working for a least squares (LS) solution in bilevel image restoration. Under the EHE criterion the network can overcome the instability and converge to a solution extremely close to the LS one. Simulation results compared with those of the ordinary Hopfield network and the simulated annealing method are presented.
In the paper we present a Hopfield neural network approach to blind bilevel image restoration. In the approach two kinds of Hopfield neural networks are used. One is the analog Hopfield neural network utilized to esti...
详细信息
暂无评论