Papers by Keyword: Texture Image

Abstract: According to the problem that classical graph-based image segmentation algorithms are not robust to segmentation of texture image. We propose a novel segmentation algorithm that GBCTRS, which overcame the shortcoming of existed graph-based segmentation algorithms N-cut and EGBIS. It extract feature vector of blocks using color-texture feature, calculate weight between each block using the neighborhood relationship, use minimum spanning tree method to clustering segmentation. The experimental show that the new algorithm is more efficient and robust to segment texture image and strong edges image.

Abstract: In image processing, the texture image segmentation is one of the most important issues. Considering the problem that the traditional segmentation methods often fail to the low quality texture image segmentation, this paper proposes a modified OTSU thresholding segmentation method. Experimental results show that the proposed method not only is well adapt to the change of brightness and contrast, but also can be applied to much complex background.

Abstract: Texture Information is widely used as one of the main low-layer features in the content-based image retrieval. In general, when the retrieval is carried out in texture image space, the same description method is adopted to regular and irregular texture images. As a large amount of regular and irregular texture images existed in the image database, it is very difficult to describe every texture with the same description method. In this paper, a retrieval strategy for texture image is proposed. The proposed strategy is divided into steps: First, classify texture images by Wold decomposition into regular and irregular texture images, then describe and retrieve them by regular and irregular texture description separately. Experimental results have showed that proposed strategy can improve classification and retrieval precision.

Abstract: Aiming at the problems of complicated convolution process of traditional wavelet transform and the unsatisfied effect of SPIHT algorithm for texture image compression, an improved SPIHT algorithm for texture image compression is proposed. At first, the texture image is decomposed into N order with the help of the lifting wavelet and the first-order high frequency sub-bands are decomposed into N-1 order by the lifting wavelet, and then the wavelet coefficients are encoded by the improved SPIHT algorithm. The improved SPIHT algorithm improved the process method of the wavelet coefficients in the low-frequency sub-bands and the detection method of some important coefficient in the L collection of the original SPIHT algorithm. Experiments show that the improved algorithm can retain the texture information of texture image more effectively and the quality of reconstructed image and peak signal to noise ratio are better than the original algorithm at the same rate. The effect is better especially at low rate, so the improved algorithm is an efficient compression method for texture image.

Abstract: In order to making the virtual experiment results close to or equivalent to the real environmental effects, in addition to model realistic scenes, also need more accurate dynamic model. In this paper, based on research of hydraulic steering paddy and wheat combine harvester, first, used three-dimensional geometric model transformation technology to convert Pro/E model to a identified geometric model in VP virtual reality systems, second, used classical dynamics equations to build the mathematical model of this vehicle dynamics, finally, used VC++ programs to connect this two models, constructed a behavioral model of complete physics properties. This behavior model is not only more realistic than the geometry model in motion, but also it follows the laws of physics movement. It can effectively strengthen the authenticity of virtual experiment, provide an important platform to start agricultural machinery virtual characteristics test, and this Modeling Method can provide a reference for other agricultural machinery virtual reality platform to build.

Abstract: Contourlet-1.3 transform has better performance in directional information representation than the original contourlet transform due to less artifacts and local frequency characteristics, and has been studied by us in retrieval systems and has been shown it is superior to contourlet ones at retrieval rate. In order to improve the retrieval rate further, a novel contourlet-1.3 transform based texture image retrieval system was proposed in this paper. In the system, sub-bands energy, standard deviation and kurtosis in contourlet domain were cascaded to form feature vectors, and the similarity measure function was Canberra distance. Experimental results show that this contourlet-1.3 transform based image retrieval system has higher retrieval rates about 7% to that of the contourlet transform with absolute mean sub-bands energy and standard deviations features under same system structure.

Abstract: Aiming at the defect that BP neural network classification model takes a long time for network training and the condition that wavelet network model lacks of direction information description, the paper presents a method for SAR image classification based on Brushlet and self-adaptive ridgelet neural network. The method extracts the energy and phase feature of SAR image texture through the Brushlet transformation, and inputs the feature vector that describes the energy and phase information into self-adaptive ridgelet neural network for training and classification. The contrast experiment indicates that the classification method proposed in this paper is rapid and accurate, and outperforms the traditional methods.

Abstract: Contourlet transform has better performance in directional information representation than wavelet transform and has been studied by many researchers in retrieval systems and has been shown that it is superior to wavelet ones at retrieval rate. In order to improve the retrieval rate further, a contourlet-S transform based texture image retrieval system was proposed in this paper. In the system, the contourlet transform was constructed by non-subsampled Laplacian pyramid cascaded by critical subsampled directional filter banks, sub-bands absolute mean energy and kurtosis in contourlet-S domain are cascaded to form feature vectors, and the similarity metric is Canberra distance. Experimental results on 109 brodatz texture images show that using the features cascaded by absolute mean and kurtosis can lead to a higher retrieval rate than the combination of standard deviation and absolute mean which is most commonly used today under same dimension of feature vectors. contourlet-S transform based image retrieval system is superior to those of the original contourlet transform, non-subsampled contourlet system under the same system structure with same length of feature vectors, retrieval time and memory needed, contourlet-S decomposition structure parameters can make significant effects on retrieval rates, especially scale number.

Abstract: Contourlet transform and wavelet have been widely used in image processing systems including texture image retrieval applications, and many literatures have reported how to construct the retrieval systems. Among them, generalize Gaussian distribution (GGD) model is a promising one. We will compare the algorithm with another one, which uses energy and standard deviation features and Canberra distance. Experimental results on 40 texture images from MIT vision database show that the latter one has higher retrieval rates for wavelet and contourlet transform retrieval systems, which indicate that the GGD model is not accurate for charactering texture feature.

Abstract: Contourlet transform is better in direction information representation than wavelet transform which has been studied in retrieval systems and has been shown that it is superior to wavelet ones at retrieval rate. In order to improve the retrieval rate further, an anti-aliasing contourlet-S transform based texture image retrieval system was proposed. In this system, the contourlet transform was constructed by anti-aliasing non-subsampled Laplacian pyramid cascaded by critical sub-sampled directional filter banks, sub-bands energy and standard deviations in contourlet domain are cascaded to form feature vectors, and the similarity metric used here is Canberra distance. Experimental results show that contourlet-S transform based image retrieval system is superior to those of the original contourlet transform, and non-subsampled contourlet system under the same system structure with almost same dimension of feature vectors, retrieval time and memory needed; and contourlet decomposition structure parameters can make significant effects on retrieval rates, especially scale number. To improve the retrieval rate of this system, kurtosis in each sub-band coefficients can be incorporated in features at the cost of some higher dimension of feature vectors.