Contourlet

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

Fusion of multisensor images using Nonsubsampled Contourlet Transform and Fuzzy Logic

Abstract: This paper presents a new fusion methodology for combining images obtained from multiple cameras using Non-Subsampled Contourlet Transform (NSCT) and Fuzzy Logic. Recent studies confirm nonsubsampled contourlet transform as an efficient tool for image fusion, as it can effectively capture inherent geometric structures in images. The proposed technique uses Non-Subsampled Contourlet Transform for multiresolution decomposition of the input images and fuzzy logic for computing the optimal fusion weights of the low-pass NSCT coefficients. The NSCT low-pass coefficients are weighted based on the activity level measurements made, in order to generate fused images of superior quality. The high frequency directional coefficients are fused, based on the local energy feature. Both pixel and region based fusion are performed using fuzzy logic and applied to multisensor images of different types. Experimental results obtained reveal that the fuzzy based fusion algorithms proposed achieve better performance when compared to the state-of-the-art fusion schemes such as dual tree complex wavelet transform (DT-CWT), NSCT and hybrid NSCT in terms of the various fusion quality evaluation metrics.

A Contourlet Domain Image Denoising Method Based on Mathematical Morphology

Abstract: A Contourlet domain image denoising method is proposed based on mathematical morphology.By using Contourlet Transform,the noised image is decomposed into a low frequency subband and a set of multisacle and multidirectional high frequency subbands.The high frequency coefficients of the original image are processed by mathematical morphological operator.The noise which have small or no at all support area are removed,and the small features which have large or consecutive support area are preserved.The denoising image will be gotten by performing the inverse Contourlet Transform to these estimated coefficients.Experimental results show that the denoising effect of this proposed method is better than that of other methods based on Contourlet transform.

Image contrast enhancement by contourlet transform and PCNN

Abstract: We put forward a contourlet transform and PCNN based image enhancement algorithm in this paper. This method overcomes the traditional wavelet transform's deficiencies in sparse representing inseparable semantic details of images, such as multidirectional edges. Firstly the images are decomposed into multi-directional and multi-scale contour segments by the contourlet transform, the directions of the line-shaped discontinuity in the images are detected. Secondly, a new physiological enhancement function based on the pulse coupled neural networks (PCNN) is proposed to enhance the coefficients by contourlet transform. At last, we reconstruct the enhanced coefficients and obtain the enhanced results of the original images, in order to verify the effectiveness of this method, we compare it to conventional contrast enhancement methods, such as the histogram equalization. The results show that our contrast enhancement method is better than other in enhancing the directional line-shaped objects.

Brain mr image segmentation by modified active contours and contourlet transform

Abstract: Multiresolution analysis is often used for image representation and processing because it can represent image at the split resolution and scale space. In this paper, a novel medical image segmentation algorithm is proposed that combines contourlet transform and modified active contour model. This method has a new energy formulation by representing the image with the coefficients of a contourlet transform. This results fast and accurate convergence of the contour towards the object boundary. Medical image segmentation using contourlet transforms has shown significant improvement towards the weak and blurred edges of the Magnetic Resonance Image (MRI). Also, the computational complexity is less compared to using traditional level sets and variational level sets for medical image segmentation. The special property of the contourlet transform is that, the directional information is preserved in each sub-band and is captured by computing its energy. This energy is capable of enhancing weak and complex boundaries in details. Performance of medical image segmentation algorithm using contourlet transform is compared with other deformable models in terms of various performance measures.

Compression and quality evaluation method for region of interest (ROI) of CT (Computed Tomography) image

Abstract: The invention provides a compression and quality evaluation method for the region of interest (ROI) of a CT (Computed Tomography) image. The method comprises ROI extraction, ROI compression and quality evaluation. The method comprises the following steps of defining the ROI according to CT image characteristics and human vision characteristics; extracting the ROI based on an image segmentation principle; marking the ROI and a non-ROI separately by adopting a MAXSHIFT algorithm; performing layered compressed encoding on the ROI based on DWT (Discrete Wavelet Transform) and EBCOT (Embedded Block Coding with Optimized Truncation); computing a structural similarity index matrix (SSIM) based on a human vision characteristic contrast sensitivity function (CSF) and contourlet transform (CT); and verifying ROI compressed image quality through CT-SSIM. The image segmentation principle is applied to the extraction of the ROI, so that the ROI comprising internal and external outline information can be extracted automatically and accurately, ROI compression is performed on a CT medical image, and a file is compressed while medical diagnosis information is kept. Secondly, through an objective quality evaluation method based on the human vision characteristics, a human vision evaluation result is approached to the maximum extent, and subjective evaluation can be replaced.