Nonsubsampled Contourlet Research Articles

Range-gated imaging in turbid conditions using a combination of intensity and polarization information

Polarization of light is introduced into range-gated imaging (RGI) for target detection and recognition in turbid conditions. The polarimetric RGI system can obtain both intensity and polarization information on objects. The experimental results show that an intensity-based image, when the target part and the background part of an object have similar reflective characteristic, can reveal the contours of the object by minimizing the effect of backscattering from the medium, and a polarization-based image can highlight the target part by common-mode rejection of the background. Furthermore, intensity and polarization components are combined to describe the scene systematically. Four methods—Laplacian pyramid, discrete wavelet transform, contourlet transform and nonsubsampled contourlet transform—are used for image fusion. Favourable information represented in each image can be extracted from image fusion for target detection and recognition. Objective evaluation indices are applied to compare the performance of different image fusion methods.

Nonsubsampled contourlet transform and k-means clustering for degraded document image binarization

Binarization is the starting step of document analysis and recognition systems. A binarization method is proposed for a degraded historical document image. The binarization methodology is based on the joint use of nonsubsampled contourlet transform (NSCT) for enhancement and k -means clustering for binarization. The input degraded image is decomposed by NSCT for generating coefficients, which are handled through a weighting scheme for highlighting significant features. The resulting reconstructed enhanced image is then binarized by mapping pixels into foreground (text) or background (no text) using k -means clustering. Experiments are conducted on document image binarization competition datasets using blind and unblind evaluation protocol. Unblind evaluation is performed on four specific types of degradations, which are stain, ink bleed-through, nonuniform background, and ink intensity variation. The obtained results show the effectiveness of the proposed scheme in terms of objective and subjective evaluations as well as stability with respect to the other well-known methods.

Block-Extraction and Haar Transform Based Linear Singularity Representation for Image Enhancement

In this paper, we develop a novel linear singularity representation method using spatial K-neighbor block-extraction and Haar transform (BEH). Block-extraction provides a group of image blocks with similar (generally smooth) backgrounds but different image edge locations. An interblock Haar transform is then used to represent these differences, thus achieving a linear singularity representation. Next, we magnify the weak detailed coefficients of BEH to allow for image enhancement. Experimental results show that the proposed method achieves better image enhancement, compared to block-matching and 3D filtering (BM3D), nonsubsampled contourlet transform (NSCT), and guided image filtering.

Zernike moments-based fingerprint recognition using weighted-support vector machine

Moments play an important role in image analysis and invariant pattern recognition. There are two types: orthogonal moments and non-orthogonal moments. Orthogonal moments perform better than non-orthogonal moments, they have properties such as robustness to image noise and geometrical invariant properties such as scale, rotation and translation. In this paper, an improvement in fingerprint recognition is done by using the Non-subsampled contourlet transform (NSCT) and the Zernike moments (ZMs). NSCT decomposes the fingerprint images into NSCT subbands. Thereafter, ZMs are used to evaluate the features of fingerprint images. Thereafter, feature selection technique is applied to select potential features from the obtained features using coefficient of determination. Thereafter, a well-known weighted-support vector machine is also used to train and test the evaluated features. Extensive experiments reveal that the proposed technique achieves significant improvement over the existing techniques in terms of accuracy, sensitivity, specificity, [Formula: see text]-measure, kappa statistics and execution time.

Retinal Vasculature Extraction using Non-Subsampled Contourlet Transform and Multi-structure Element Morphology by Reconstruction

Retinal vasculature extraction is an area of utmost interest in ophthalmology. It helps to diagnose various diseases and also play a crucial role in treatment planning and accomplishment.In this work, we suggest an algorithm to segment retinal vasculature from retinal Fundus Images(FI) using multi-structure element morphology after enhancing the image using Normal Inverse Gaussian (NIG) model in the fuzzified Non-Subsampled Contourlet Transform (NSCT) domain. Since both noises and weak edges produce low magnitude NSCT coefficients, image enhancement methods amplify weak edges as well as noises. Direct application of image boosting technique in the NSCT domain causes over enhancement. So a novel image enhancement method is employed by interpreting the term “contrast” as a qualitative instead of a quantitative measure of the image. Membership values of NSCT coefficients are modified using NIG model. Mathematical Morphology(MM) by Multi-structure Elements (MEs) is used to extract the edges of image. False vessel ridges are expunged, and the thin vessel edges are preserved using opening by reconstruction. Connected component analysis followed by length filtering is used to filter the still remaining false edges. In most of the available literature, low-resolution fundus image databases are used for evaluating the algorithm. In our work, we evaluate our algorithm not only utilizing the DRIVE database, a low-resolution retinal image (RI) database, but also using an openly available High-Resolution Fundus (HRF) image database. Our result illustrates that the proposed method outperforms the other techniques considered with average accuracy (ACC) of 96.71%. In addition to ACC, we also use F1-Score and Mathews Correlation Coefficient (MCC) to evaluate our method. The average values of the results obtained with the HRF image database for F1-Score and MCC are 0.8172 and 0.8031, respectively, which are very much encouraging.

Computer aided tool for automatic detection and delineation of nucleus from oral histopathology images for OSCC screening

Characterization of nucleus from oral tissue histology images is important for oral squamous cell carcinoma (OSCC) diagnosis. Therefore, the accurate delineation of nucleus is the foremost task to initiate the quantification of its various properties such as morphology, texture, intensity, etc. In this paper, we propose a two-stage computer-aided tool for automatic detection (the first stage) and delineation of the detected nucleus (second stage) from oral histological images to aid clinician in OSCC diagnosis. In the first stage, 81×81 patches were extracted from our image database followed by the wavelet-based downsampling process to generate 21×21 patches, which were used to train the deep convolution neural network (CNN) architecture. This approach was used to automatically detect nucleus from OSCC images and achieved 88.87% recall and 82.03% precision. After detection, 81×81 image patches were extracted from those detected nucleus regions and further processed through the second stage of proposed pipeline to segment the exact boundary of nucleus using active contour (Chan-Vese model) after applying edge enhancing nonsubsampled contourlet transform. The proposed segmentation methodology performed well with 94.22% Dice coefficient, 89.38% Jaccard index, 97.56% precision, and 91.58% recall. We view this work as the first attempt on oral tissue histology image computation for joint nucleus detection and segmentation to diagnose OSCC to the best of our knowledge.

Coefficient difference based watermark detector in nonsubsampled contourlet transform domain

Imperceptibility and robustness are two main requirements of any image watermarking systems to guarantee desired functionalities, but there is a tradeoff between them from the information–theoretic perspective. It is a challenging work to design a high performance digital watermarking scheme to keep a trade-off between imperceptibility and robustness. By modeling the nonsubsampled Contourlet transform (NSCT) coefficient differences with ranked set sample (RSS) based Cauchy distribution and employing locally most powerful (LMP) test, we propose a locally optimum image watermark detector in NSCT domain. In the proposed scheme, we first compute the robust coefficient differences according to the inter-scale dependency between NSCT coefficients, and then embed the digital watermark into the significant NSCT coefficient difference subband. At the watermark receiver, robust NSCT coefficient differences are firstly modeled by employing the Cauchy distribution, where the statistical properties of NSCT coefficient differences are captured accurately. Then the RSS approach is introduced to estimate the location parameter and shape parameter of Cauchy distribution. And finally an optimal detector for multiplicative watermarking is developed using the LMP decision rule and RSS-based Cauchy distribution. Also, we utilize the statistical model to derive the closed-form expressions for the watermark detector. Experimental results demonstrate the high efficiency of our watermarking scheme, which can provide better imperceptibility and outstanding robustness against various attacks, in comparison with the state-of-the-art approaches recently proposed in the literature.

An Effective Method for Image Denoising Using Non-local Means and Statistics based Guided Filter in Nonsubsampled Contourlet Domain

An Effective Method for Image Denoising Using Non-local Means and Statistics based Guided Filter in Nonsubsampled Contourlet Domain

Electromagnetic Induction Heating and Image Fusion of Silicon Photovoltaic Cell Electrothermography and Electroluminescence

In the process of research, development, production, service, and maintenance of silicon photovoltaic (Si-PV) cells and the requirements for detection technology are becoming more and more important. This paper aims to investigate electromagnetic induction (EMI) and image fusion to improve the detection effect of electrothermography (ET) and electroluminescence (EL) of multidefects in Si-PV cells. First, the principles of ET, EL, and other physical processes including EMI, thermal radiation, and luminescence radiation are analyzed in this paper. ET and EL techniques after EMI improvement are used to detect different defects including scratch, broken gridline, surface impurity, hidden crack, and so on. The qualitative results show that EMI can greatly improve the defect detection ability of ET and EL. Then, an image-fusion rule based on L1 norm is proposed to fuse the sparse vector of the ET and EL images. The integration and complementarity of the two wavelength detection data are achieved. Finally, the image-fusion results of sparse representation (SR) algorithm is compared with discrete wavelet transform, curvelet transform, dual-tree complex wavelet transforms, and nonsubsampled contourlet transform. Five objective evaluation indexes including root mean square error, peak signal-to-noise ratio, correlation coefficient, mutual information, and structural similarity index are used to evaluate the fusion results. Overall evaluation results show that the SR algorithm is superior to the other algorithms.

Multimodel fusion method via sparse representation at pixel-level and feature-level

In current fusion methods based on sparse representation (SR) and different frequency, the SR is usually applied to fusion of the low-frequency components. In contrast, the direct fusion is usually adopted for high-frequency components due to their significant diversity. However, the effect of the direct fusion is degraded by the redundant information resulting from the correlation between original signals. A multimodel fusion framework is proposed by applying the SR to low-frequency fusion at pixel-level and high-frequency fusion at feature-level, respectively. First, the multimodal images are decomposed into high-frequency and low-frequency components by nonsubsampled contourlet transform (NSCT). Second, the universal high-frequency dictionary is constructed by using the fast independent component analysis (ICA) of the source high-frequency and its subband images. They represent the general feature part and unique feature part for the high-frequency signals, respectively. The universal low-frequency dictionary is constructed by using the original low-frequency signals. Third, the direct fusion of the high-frequency is converted into sparse coefficients fusion in fast ICA domain. Moreover, the multiple directive contrasts by modifying sum-modified Laplacian are taken as the fusion rule. The low-frequency signals are fused by using an activity measure based on weights. Finally, the fused image is obtained by inverse NSCT on the merged components. The experiments are conducted on three types of image pairs, and the results demonstrate that the proposed method outperforms seven state-of-the-art methods, in terms of four subjective and objective evaluations.

MRI/CT fusion based on latent low rank representation and gradient transfer

Medical image fusion mainly focuses on finding better method on combining multi-modal medical images with different characteristics, and which has been playing a significant role on clinical diagnosis and disease treatment. For the fusion of MRI (Magnetic Resonance Imaging) and CT (Computed Tomography), a novel algorithm is proposed in this paper based on LatentLRR (Latent Low Rank Representation) and gradient transfer. LatentLRR is a meaningful tool for separating out detailed information and saliency information. In order to produce saliency information fused result, image statics is used through the corresponding detailed information. For detailed information, it is separated into high frequency parts and low frequency parts by NSCT (Non-Subsampled Contourlet Transform), the former are fused by the method of scale based, and the latter are merged by local energy maxima. To obtain the reconstructed image, it needs to integrate high frequency fused parts and low frequency fused parts into one image by inverse NSCT. Finally, the final image can be given by combining the reconstructed images and saliency information fused result into one image, and for a better visual effect, optimize the final result by gradient transfer. Compared with the state-of-the-art methods on the experiments of ten pair clinical medical images MRI/CT, the proposed algorithm receives a comprehensive advantage in preserving the detailed and gradient information, not only in visual effects but also in objective evaluation.

Locally optimum image watermark decoder by modeling NSCT domain difference coefficients with vector based Cauchy distribution

Improving the ability of imperceptibility, watermark capacity, and robustness at the same time still remains a challenge within the digital image watermarking community. By modeling the robust nonsubsampled Contourlet transform (NSCT) difference coefficients with vector based Cauchy distribution and employing locally most powerful (LMP) test, we propose a locally optimum image watermark decoder in NSCT domain. We first compute the difference coefficients according to the inter-scale dependency between NSCT coefficients, and investigate the robustness of the NSCT difference coefficients by subjective visual error and objective mean squared error (MSE) terms. We then embed the digital watermark into the significant NSCT difference subband with highest energy by modifying the robust NSCT difference coefficients. At the receiver, by combining the vector based Cauchy probability distribution and LMP test, we propose a locally optimum blind watermark decoder in the NSCT domain. Here, robust NSCT difference coefficients are firstly modeled by employing the vector based Cauchy probability density function (PDF), where the Cauchy marginal statistics and various strong dependencies of NSCT coefficients are incorporated. Then the statistical model parameters of vector based Cauchy PDF are estimated using second-kind statistics approach. And finally a blind image watermark decoder is developed using vector based Cauchy PDF and LMP decision rule. We conduct extensive experiments to evaluate the performance of the proposed blind watermark decoder, in which encouraging results validate the effectiveness of the proposed technique, in comparison with the state-of-the-art approaches recently proposed in the literature.

A Practical Medical Image Enhancement Algorithm Based on Nonsubsampled Contourlet Transform

A Practical Medical Image Enhancement Algorithm Based on Nonsubsampled Contourlet Transform

De-scattering and edge-enhancement algorithms for underwater image restoration

Image restoration is a critical procedure for underwater images, which suffer from serious color deviation and edge blurring. Restoration can be divided into two stages: de-scattering and edge enhancement. First, we introduce a multi-scale iterative framework for underwater image de-scattering, where a convolutional neural network is used to estimate the transmission map and is followed by an adaptive bilateral filter to refine the estimated results. Since there is no available dataset to train the network, a dataset which includes 2000 underwater images is collected to obtain the synthetic data. Second, a strategy based on white balance is proposed to remove color casts of underwater images. Finally, images are converted to a special transform domain for denoising and enhancing the edge using the non-subsampled contourlet transform. Experimental results show that the proposed method significantly outperforms state-of-the-art methods both qualitatively and quantitatively.

Denoising Method for Passive Photon Counting Images Based on Block-Matching 3D Filter and Non-Subsampled Contourlet Transform.

Multi-pixel photon counting detectors can produce images in low-light environments based on passive photon counting technology. However, the resulting images suffer from problems such as low contrast, low brightness, and some unknown noise distribution. To achieve a better visual effect, this paper describes a denoising and enhancement method based on a block-matching 3D filter and a non-subsampled contourlet transform (NSCT). First, the NSCT was applied to the original image and histogram-equalized image to obtain the sub-band low- and high-frequency coefficients. Regional energy and scale correlation rules were used to determine the respective coefficients. Adaptive single-scale retinex enhancement was applied to the low-frequency components to improve the image quality. The high-frequency sub-bands whose line features were best preserved were selected and processed using a symbol function and the Bayes-shrink threshold. After applying the inverse transform, the fused photon counting image was subjected to an improved block-matching 3D filter, significantly reducing the operation time. The final result from the proposed method was superior to those of comparative methods in terms of several objective evaluation indices and exhibited good visual effects and details from the objective impression.

Infrared Dim Target Detection Using Shearlet’s Kurtosis Maximization under Non-Uniform Background

A novel method based on multiscale and multidirectional feature fusion in the shearlet transform domain and kurtosis maximization for detecting the dim target in infrared images with a low signal-to-noise ratio (SNR) and serious interference caused by a cluttered and non-uniform background is presented in this paper. First, an original image is decomposed using the shearlet transform with translation invariance. Second, various directions of high-frequency subbands are fused and the corresponding kurtosis of fused image is computed. The targets can be enhanced by strengthening the column with maximum kurtosis. Then, processed high-frequency subbands on different scales of images are merged. Finally, the dim targets are detected by an adaptive threshold with a maximum contrast criterion (MCC). The experimental results show that the proposed method has good performance for infrared target detection in comparison with the nonsubsampled contourlet transform (NSCT) method.

Image speckle noise denoising by a multi-layer fusion enhancement method based on block matching and 3D filtering

ABSTRACTIn order to improve speckle noise denoising of block matching and 3D filtering (BM3D) method, an image frequency-domain multi-layer fusion enhancement method (MLFE-BM3D) based on nonsubsampled contourlet transform (NSCT) has been proposed. The method designs an NSCT hard threshold denoising enhancement to preprocess the image, then uses fusion enhancement in NSCT domain to fuse the preliminary estimation results of images before and after the NSCT hard threshold denoising, finally, BM3D denoising is carried out with the fused image to obtain the final denoising result. Experiments on natural images and medical ultrasound images show that MLFE-BM3D method can achieve better visual effects than BM3D method, the peak signal to noise ratio (PSNR) of the denoised image is increased by 0.5 dB. The MLFE-BM3D method can improve the denoising effect of speckle noise in the texture region, and still maintain a good denoising effect in the smooth region of the image.

Interval type-2 fuzzy set and human vision based multi-scale geometric analysis for text-graphics segmentation

This paper presents a novel method for texture-based text-graphic segmentation in a text embedded image. In the method, features are computed applying Multi-scale Geometric Analysis(MGA). The MGA of the image is done by Nonsubsampled contourlet transform(NSCT). The NSCT sub-bands help to generate the features which represent textures of the text portions and graphics portions of the image. In a segmentation process, the uncertainties arise mainly for two reasons: one is the ambiguity in gray level and other is the spatial ambiguity. Here the uncertainties are managed by interval type2 fuzzy set (IT2FS). The human vision model called human psychovisual phenomenon (HVS) is incorporated in the process for generating the interval type-2 fuzzy membership functions (IT2FMF). The efficiency of the proposed scheme is measured on the benchmark dataset. The robustness and performance bound of the proposed technique under noise corruption are measured statistically using modified Cramer-Rao bound. We found that effectiveness of the features by NSCT in combination with the IT2FS are quite promising in comparison to the state-of-the-arts methods.

Image classification with an RGB-channel nonsubsampled contourlet transform and a convolutional neural network

In this paper, an efficient image classification method is proposed that is based on the nonsubsampled contourlet transform (NSCT) of RGB-channel images and the convolutional neural network (CNN). First, the NSCT-based coefficients of natural RGB-channel images are extracted, which are capable of capturing the statistical properties of each channel. In addition, the proposed feature descriptor is equipped with the mean–max-pooling strategy according to the characteristics of the correlated coefficients. Then, the CNN is concatenated to exaggerate the discriminative parts of the primary features. With these advantages, the proposed RGB-channel NSCT–CNN should, in general, improve the corresponding CNN-based image classification methods. Using the Food-101 and SUN Datasets, the proposed method achieves state-of-the-art classification results that are also significant for object detection. In addition, the proposed method can achieve better or comparable accuracy compared to other related methods based on these two datasets.

Weighted image fusion using cross bilateral filter and non-subsampled contourlet transform

This paper provides a fusion technique for multi-focus imaging using cross bilateral filter and non-subsampled contourlet transform. The snapshots are decomposed into distinct approximation and detail components. The original image and approximation component is passed through cross bilateral filter to obtain approximation weight map. Whereas, the detail components are combined using weighted average to obtain detail weight map. The weights are combined together and used with original images to obtain the resultant fused image. Visual and quantitative analysis shows the significance of proposed scheme in comparison to state of art fusion schemes.