scispace - formally typeset
Search or ask a question
Topic

Contourlet

About: Contourlet is a research topic. Over the lifetime, 3533 publications have been published within this topic receiving 38980 citations.


Papers
More filters
Proceedings ArticleDOI
07 May 2001
TL;DR: An efficient filter bank for reconstruction in the Laplacian pyramid is proposed that is shown to perform better than the usual method and indicate that gains of more than 1 dB are actually achieved.
Abstract: We study the Laplacian pyramid (LP) as a frame operator, and this reveals that the usual reconstruction is suboptimal. With orthogonal filters, the LP is shown to be a tight frame, thus the optimal linear reconstruction using the dual frame operator has a simple structure as symmetrical with the forward transform. For more general cases, we propose an efficient filter bank for reconstruction in the LP that is shown to perform better than the usual method. Numerical results indicate that gains of more than 1 dB are actually achieved.

20 citations

Proceedings ArticleDOI
01 Jul 2007
TL;DR: Experimental results demonstrate that the SVFNN method presented is a promising method for mass classification in mammography.
Abstract: In this paper, a new approach for mass classification in digital mammograms based on contourlet texture features and support-vector-based fuzzy neural network (SVFNN) classifier is presented. The SVFNN combines the superior classification power of support vector machine (SVM) in high dimensional data spaces, the efficient human-like reasoning of fuzzy in handling uncertainty information, and learning property of neural networks. Each mammogram is segmented to regions of interest and features are extracted in frequency domain by contourlet coefficients. One of the main contribution of this research is taking benefit from the superiority of the contourlet compared to the multi-scale techniques and use SVFNN for mass classification. MIAS1 data set is used to evaluate the proposed method. Experimental results demonstrate that the method presented is a promising method for mass classification in mammography.

19 citations

18 May 2009
TL;DR: This report develops the discrete shearlet transform (DST) which provides efficient multiscale directional representation and shows that the implementation of the transform is built in the discrete framework based on a multiresolution analysis.
Abstract: It is now widely acknowledged that analyzing the intrinsic geometrical features of an underlying image is essentially needed in image processing. In order to achieve this, several directional image representation schemes have been proposed. In this report, we develop the discrete shearlet transform (DST) which provides efficient multiscale directional representation. We also show that the implementation of the transform is built in the discrete framework based on a multiresolution analysis. We further assess the performance of the DST in image denoising and approximation applications. In image approximation, our adaptive approximation scheme using the DST significantly outperforms the wavelet transform (up to 3.0dB) and other competing transforms. Also, in image denoising, the DST compares favorably with other existing methods in the literature.

19 citations

Journal ArticleDOI
TL;DR: Improved version of the first adaptive image steganography methods in which it uses an advanced embedding operation to boost its security has been proposed, and has superior performance in terms of PSNR and SSIM, and is more secure against the steganalysis attack.
Abstract: This work presents adaptive image steganography meth-ods which locate suitable regions for embedding by contourlet trans-form, while embedded message bits are carried in discrete cosinetransform coefficients The first proposed method utilizes contourlettransform coefficients to select contour regions of the image In theembedding procedure, some of the contourlet transform coefficientsmay change which may cause errors at the message extractionphase We propose a novel iterative procedure to resolve such prob-lems In addition, we have proposed an improved version of the firstmethod in which it uses an advanced embedding operation to boostits security Experimental results show that the proposed basemethod is an imperceptible image steganography method with zeroretrieval error rate Comparisons with other steganography methodswhich utilize contourlet transform show that our proposed method isable to retrieve all messages perfectly, whereas the others failMoreover, the proposed method outperforms the ContSteg methodin termsofPSNRandthehigher-order statisticssteganalysismethodExperimental evaluations of our methods with the well known DCT-based steganography algorithms have demonstrated that ourimproved method has superior performance in terms of PSNR andSSIM, and is more secure against the steganalysis attack © 2013SPIE and IS&T [DOI: 101117/1JEI224043007]

19 citations

Proceedings ArticleDOI
18 Dec 2006
TL;DR: The proposed method is compared with a wavelet based blind technique and the results prove that contourlet based technique gives better robustness, under similar embedding conditions.
Abstract: This paper presents a novel method of blind image watermarking in contourlet domain. We have used spread spectrum technique for additive watermark embedding. A correlation detector is used to detect the embedded pseudorandom sequence. The binary logo thus retrieved proves authenticity of the image. The similarity of the retrieved binary logo with the original embedded logo is veriJied using correlation technique. Post processing of the retrieved logo gives better visual effects, further aiding threshold selection for detection. We have verijied the robustness of the proposed method against dzrerent attacks including StirMark attack. The proposed method is compared with a wavelet based blind technique and the results prove that contourlet based technique gives better robustness, under similar embedding conditions.

19 citations


Network Information
Related Topics (5)
Feature extraction
111.8K papers, 2.1M citations
89% related
Image processing
229.9K papers, 3.5M citations
85% related
Convolutional neural network
74.7K papers, 2M citations
84% related
Deep learning
79.8K papers, 2.1M citations
82% related
Artificial neural network
207K papers, 4.5M citations
81% related
Performance
Metrics
No. of papers in the topic in previous years
YearPapers
202336
202299
202175
2020109
2019155
2018164