Showing papers on "Contourlet published in 2004"
24 Oct 2004
TL;DR: A new coding technique based on a mixed contourlet and wavelet transform that is optimized through an iterative projection process in the transform domain in order to minimize the quantization error in the image domain is presented.
Abstract: This paper presents a new coding technique based on a mixed contourlet and wavelet transform. The redundancy of the transform is controlled by using the contourlet at fine scales and by switching to a separable wavelet transform at coarse scales. The transform is then optimized through an iterative projection process in the transform domain in order to minimize the quantization error in the image domain. A gain of respectively up to 0.5 dB and to 1 dB over respectively contourlet and wavelet based coding has been observed for images with directional features.
43 citations
Journal Article•
TL;DR: An image denosing algorithm using Curvelet transform free of 搘arp-around?artifacts, but also combined wavelet thresholding with edge preserving approach to retain the sharpness of the images.
Abstract: An image denosing algorithm using Curvelet transform is proposed. We not only adopted an improved Curvelet transform free of 搘arp-around?artifacts, but also combined wavelet thresholding with edge preserving approach to retain the sharpness of the images. Experiments show that our method yields denoised images with higher PSNR value and better visual quality. Key words: image processing; image denoising; Curvelet transform
15 citations
01 Jan 2004
TL;DR: A novel method of Synthetic Aperture Radar image despeckling using the contourlet transform representation is presented and it is shown that thecontourlet methods have better performance than multilook or wavelet methods.
Abstract: A novel method of Synthetic Aperture Radar (SAR) image despeckling using the contourlet transform representation is presented. Justification for the use of the contourlet signal representation, originally developed for natural images, is given. Methods of evaluating the despeckling performance of various algorithms are provided. Finally, a comparison of performance of multilook processing, wavelet-based despeckling, contourlet- based despeckling, and contourlet-based despeckling with cycle spinning is provided for both simulated and actual SAR images. It is shown that the contourlet methods have better performance than multilook or wavelet methods.
11 citations
TL;DR: A novel approach is presented that combines two types of features extracted by discrete wavelet transform and contourlet transform, and shows that the combined features result in better classification rates than using only one type of those.
Abstract: In the recent decades, many features used to represent a texture were proposed. However, these features are always used exclusively. In this paper, a novel approach is presented that combines two types of features extracted by discrete wavelet transform and contourlet transform. Support vector machines (SVMs), which have demonstrated excellent performance in a variety of pattern recognition problems, are used as classifiers. The algorithm is tested on four different datasets, selected from Brodatz and VisTex database. The experimental results show that the combined features result in better classification rates than using only one type of those.
8 citations
20 Dec 2004
TL;DR: In this article, a block based texture segmentation method based on contourlets and the hidden Markov model (HMM) is proposed, which combines HMM-HMT model that models global dependency between the blocks in addition to the local statistics within a block.
Abstract: In this paper, block based texture segmentation is proposed based on contourlets and the hidden Markov model (HMM). Hidden Markov model is combined with hidden Markov tree (HMT) to form HMM-HMT model that models global dependency between the blocks in addition to the local statistics within a block. The HMM-HMT model is modified to use the contourlet transform, a new extension to the wavelet transform that forms a true basis for image representations. The maximum likelihood multiresolution segmentation algorithm is used to handle several block sizes at once. Since the algorithm works on the contourlet transformed image data, it can directly segment images without the need for transforming into the space domain. The experimental results demonstrate the competitive performance of the algorithm on contourlets with that of the other methods and excellent visual performance at small block sizes.The performance is comparable with that of wavelets and is superior at small block sizes.
8 citations
DOI•
01 Jan 2004
TL;DR: The results suggest that these directional-selective basis functions provide a usefull tool for the removal of coherent noise such as ground roll.
Abstract: In this paper we present examples of ground roll attenuation for synthetic and real data gathers by using Contourlet and Curvelet transforms. These non-separable wavelet transforms are locoalized both (x,t)and (k,f)-domains and allow for adaptive seperation of signal and ground roll. Both linear and non-linear filtering are discussed using the unique properties of these basis that allow for simultaneous localization in the both domains. Eventhough, the linear filtering techniques are encouraging the true added value of these basis-function techniques becomes apparent when we use these decompositions to adaptively substract modeled ground roll from data using a non-linear thesholding procedure. We show real and synthetic examples and the results suggest that these directional-selective basis functions provide a usefull tool for the removal of coherent noise such as ground roll.
8 citations
26 Jul 2004
TL;DR: A directional image for a fingerprint obtained from an expanded second stage Haar wavelet transform is presented, which represent the orientation pattern that later on are used for fingerprint matching so that the real owner can be identified.
Abstract: A directional image for a fingerprint obtained from an expanded second stage Haar wavelet transform is presented. This directional image represent the orientation pattern that later on are used for fingerprint matching so that the real owner can be identified. To construct the directional image, we first perform a second stage Haar wavelet transform on a fingerprint. Due to the limitation of the Haar wavelet, we use the expanded version of it instead so that a better directional image can be obtained. Next, we quantize the directional image into a few grey-level values that represent a range of angle orientations. We then smooth the image using an averaging filter. Based on this image we then perform the matching process based on the minimum square error (MSE) criteria.
7 citations
Journal Article•
TL;DR: The paper explained the method of image fusion based on laplacian pyramid and contrast pyramid and used these methods for multi- focus image and remote sensing image fusion and compared the quality of these fusion images.
Abstract: The paper explained the method of image fusion based on laplacian pyramid and contrast pyramid.Then use these methods for multi- focus image and remote sensing image fusion.Also this paper compared the quality of these fusion images,then obtain characters of two methods.
6 citations
01 Aug 2004
TL;DR: A newly proposed multiresolution non-uniform directional filter bank (nuDFB) is applied to the problem of texture classification and shows that the nuDFB offers a better representation of texture with directional edges.
Abstract: A newly proposed multiresolution non-uniform directional filter bank (nuDFB) is applied to the problem of texture classification. The classification performance is compared with other critically sampled filter banks, including the conventional directional filter bank and the discrete wavelet transform, and shows that the nuDFB offers a better representation of texture with directional edges.
6 citations
25 Jul 2004
TL;DR: This paper shows that the emphasis degree is changed by changing the band-width of the Gaussian filter in order to improve the performance of the enlargement method based on Laplacian pyramid representation.
Abstract: The Laplacian pyramid is the hierarchical expression. Based on Laplacian pyramid representation, the prediction of unknown higher-frequency components is equivalent to the prediction of an unknown high-resolution Laplacian image. Gaussian filter is used for calculating of the Laplacian pyramid. And the band-width of the Gaussian filter is optimal for image compression. However, we cannot assert the band-width is optimal for the image enlargement method. In this paper, we change the band-width of the Gaussian filter in order to improve the performance of the enlargement method based on Laplacian pyramid representation. We show that the emphasis degree is changed by changing the band-width of the Gaussian filter.
6 citations
01 Jan 2004
TL;DR: The authors' curvelet transform uses FRIT as a central step and implements curvelet subbands with a combined wavelet filter bank and enjoys a competitive denoising capability and surprising visual performances shown by numerical experiments.
Abstract: A digital implementation of curvelet transform is proposed which offers exact reconstruction and low redundancy (<16/3). We improved the finite ridgelet transform (FRIT) Minh N. Do and Martion Vetterli., (2003) by mains of frequency lapped orthogonal transform. Our curvelet transform uses FRIT as a central step and implements curvelet subbands with a combined wavelet filter bank. Despite the crudeness of our thresh-holding rule, the new strategy enjoys a competitive denoising capability and surprising visual performances shown by numerical experiments.
TL;DR: An overview is given on the application of directional basis functions, known under the name Curvelets/Contourlets, to various aspects of seismic processing and imaging, which involve adaptive subtraction, and applications that include multiple, ground-roll removal and migration denoising.
Abstract: In this paper an overview is given on the application of directional basis functions, known under the name Curvelets/Contourlets, to various aspects of seismic processing and imaging, which involve adaptive subtraction. Key concepts in the approach are the use of (i) directional basis functions that localize in both domains (e.g. space and angle); (ii) non-linear estimation, which corresponds to localized muting on the coefficients, possibly supplemented by constrained optimization. We will discuss applications that include multiple, ground-roll removal and migration denoising.
20 Oct 2004
TL;DR: It is found that directional analysis is suitable for characterizing structured textures, but not random textures, so structured and random textures are separated by employing an entropy-based measure on the multiscale directional features.
Abstract: The use of multiscale directional decomposition, achieved by combining a Laplacian pyramid and a directional filter bank, is studied for texture classification. We first demonstrate the importance of the multiscale analysis of directional texture features. Then, it is found that directional analysis is suitable for characterizing structured textures, but not random textures. Thus, structured and random textures are separated by employing an entropy-based measure on the multiscale directional features. Through this pre-filtering step, structured textures are extracted for further classification, so that the overall retrieval performance can be enhanced. Experimental results showed that this pre-filtering step can significantly improve the overall retrieval accuracy.
15 Jun 2004
TL;DR: The analysis shows that directional wavelet transform can better reflect the edge information of images because it better corresponds to the characteristics of image direction and texture.
Abstract: In order to process the pulp fiber image more effectively, according to its characteristics, a novel type directional wavelet transform is applied. In this paper, the relationship and differences between traditional and directional wavelet transforms are shown. Our analysis shows that directional wavelet transform can better reflect the edge information of images because it better corresponds to the characteristics of image direction and texture. The experiment proves that the directional characterizations of the fibre image can be extracted effectively.
DOI•
01 Jan 2004
TL;DR: An overview is given on the application of directional basis functions, known under the name Curvelets/Contourlets, to various aspects of seismic processing and imaging, and applications that include multiple and ground roll removal; sparseness-constrained least-squares migration and the computation of 4-D difference cubes.
Abstract: In this paper an overview is given on the application of directional basis functions, known under the name Curvelets/Contourlets, to various aspects of seismic processing and imaging. Key conceps in the approach are the use of (i) directional basis functions that localize in both domains (e.g. space and angle); (ii) non-linear estimation, which corresponds to localized muting on the coefficients, possibly supplemented by constrained optimization (iii) invariance of the basis functions under the imaging operators. We will discuss applications that include multiple and ground roll removal; sparseness-constrained least-squares migration and the computation of 4-D difference cubes.
11 Dec 2004
TL;DR: The HMM-contourlet HMT method provides superior texture segmentation results and excellent visual performance at small block sizes.
Abstract: Contourlets have emerged as a new mathematical tool for image processing and provide compact and decorrelated image representations Hidden Markov modeling (HMM) of contourlet coefficients is a powerful approach for statistical processing of natural images In this paper, we extended the hidden Markov modeling framework to contourlets and combined hidden Markov trees (HMT) with hidden Markov model to form HMM-Contourlet HMT model The model is used for block based multiresolution texture segmentation The performance of the HMM-contourlet HMT texture segmentation method is compared with that of HMM-real HMT and HMM-complex HMT methods The HMM-contourlet HMT method provides superior texture segmentation results and excellent visual performance at small block sizes
Journal Article•
TL;DR: The experiments show that the proposed approach can remove the visual artifacts, obtaining better visual result and higher PSNR.
Abstract: This paper proposes a Contourlet de-noising method based on transla tion invariance. We compare this method with the translation invariance wavelet method. The experiments show that the proposed approach can remove the visual ar tifacts, obtaining better visual result and higher PSNR.
Proceedings Article•
01 Sep 2004
TL;DR: This paper presents a tool which can be tuned relatively to image features by decomposing them into a (linear) frame of directional wavelets with variable angular selectivity and exploits some particularities of the (biorthogonal) circular multiresolution framework in the frequency domain.
Abstract: During the last ten years, many techniques have been devised to add directionality in image processing. We may cite, for instance, directional wavelets [1], steerable filters [8], curvelets [2], contourlets [5], …. This has opened the door to new ways for efficiently representing objects by oriented atoms. Real images, however, contain more than just smooth curves and straight lines defining contours of objects. They can present also details that are less oriented and more isotropic (like corners, spots, texture elements, …). We present in this paper a tool which can be tuned relatively to these image features by decomposing them into a (linear) frame of directional wavelets with variable angular selectivity. To obtain such a decomposition, these new functions exploit some particularities of the (biorthogonal) circular multiresolution framework in the frequency domain. This link suggests the name of our method, ‘multiselectivity analysis’.