Low-frequency Coefficients Research Articles

Spectral preservation fusion for remote sensing images using focus measure operators based on fast discrete curvelet transform and hyperspherical color space

How to preserve the spectral information when enhancing the spatial details is a key issue of remote sensing image fusion. The component substitution (CS)-based fusion methods can effectively enhance the spatial details while suffering spectral distortion, and multiresolution analysis (MRA)-based methods have advantages in preserving spectral information but are not satisfactory in terms of spatial details. This paper proposes a hybrid method to integrate the advantages of CS- and MRA-based approaches. The intensity image is first obtained from an original multispectral (MS) image by hyperspherical color space (HCS) transform; then, the intensity image and original panchromatic (PAN) image are decomposed by fast discrete curvelet transform (FDCT). The focus measure operators are introduced to fused low-frequency, middle-high frequency, and fine scale subband coefficients in curvelet domain with specific fusion strategies. The final fusion image is achieved by inverse FDCT and inverse HCS transform. From the aspects of subjective and objective quality assessments, the experimental results on various types of remote sensing images including IKONOS, QuickBird, and WorldView-2 indicate that compared with existing well-known algorithms and commercial softwares, the proposed method shows obvious advantages in preserving spectral information and maintaining the spatial details.

Multimodal sensor medical image fusion based on nonsubsampled shearlet transform and S-PCNNs in HSV space

Abstract Computational imaging plays an important role in medical treatment for providing more comprehensive medical images. This work proposes a new scheme to fuse computed tomography (CT), magnetic resonance (MRI), and positron emission tomography (PET) images into a single image. A novel two-stage medical image fusion scheme, which is based on non-subsampled shearlet transform (NSST) and simplified pulse coupled neural networks (S-PCNNs), is proposed in the hue-saturation-value (HSV) color space. Firstly, CT and MRI images are decomposed into a set of low and high frequency coefficients by NSST, PET images are transformed into the HSV color space, and then the V component of PET image in the HSV color space. Secondly, intersecting cortical models (ICMs) are utilized to extract the edges and outlines in a larger area from the high frequency coefficients, and S-PCNNs are employed to describe the subtly detailed information in a smaller area. Thirdly, different fusion rules are designed to fuse the corresponding low and high frequency coefficients. At last, the fused medical image is obtained by the inverse HSV and NSST transformation, successively. The experimental results show that the proposed scheme is effective, and it can fuse more information into the final images than conventional methods.

Multispectral medical image fusion scheme based on hybrid contourlet and shearlet transform domains.

Multimodal medical image sensor fusion has revolutionized the medical analysis by improving the precision of computer assisted diagnosis. This is incorporated by highlighting the complementary information while minimizing the redundant content in the fused images from various biomedical sensors like MRI, Computed Tomography, and Positron Emission Tomography/Single-Photon Emission Computerized Tomography. Multispectral image fusion is a special case of multimodal fusion which serves to encompass both spatial and spectral details in the fused image. This paper presents a hybrid sub-band decomposition scheme for multispectral image fusion comprising of non-subsampled contourlet transform and shearlet transform domains. The pre-processing stage involves color transformation of an input multispectral image from red-green-blue to YIQ color space. Thereafter, both the source images (i.e., panchromatic and multispectral images) after sub-band decomposition are processed via the application of contrast enhancement, weighted-principal component analysis, and max-max algorithms. The low frequency coefficients are processed via phase congruency whereas a combination of directive contrast and normalized Shannon entropy is applied to high frequency coefficients. The objective assessment of image quality has been carried out using various reference and no-reference based performance metrics. The distinguishing fusion response of the proposed hybrid scheme has been validated by the comparisons done with the other fusion approaches.

Infrared and visible image fusion method based on sparse features

Since the object information cannot be extracted efficiently by the traditional infrared and visible image fusion algorithms, an infrared and visible image fusion method based on the non-subsampled shearlet transform (NSST) and sparse structure features is proposed to retain the context information on visible image in this study. Firstly, we decompose the source images into low-frequency sub-band and high-frequency sub-band coefficients by the NSST. Then, benefit from the advantage of PCA on extracting principle information, the fusion rule in low-frequency sub-bands coefficients are merged by using the PCA-based approach. Afterwards, to retain the sparse structures from source images better, we propose a novel sparse feature extraction on high-frequency sub-band coefficients and fuse high-frequency components of source images. Finally, the inverse NSST is employed to obtain the fused image. The experimental results demonstrate that the proposed method preserves the background information on visible image and highlights the structural information on infrared image.

An effective image fusion algorithm based on grey relation of similarity and morphology

Considering the problems of the limited energy in wireless multi-media sensor networks (WMSNs) and the focused regions discontinuity of the fused image obtained using traditional multi-scale analysis tools (MST)-based methods, an effective multi-focus image fusion algorithm is proposed in this paper. In this method, the original fused image is obtained based on wavelet transform where the low-frequency coefficients are fused by average scheme, whereas the high-frequency coefficients are fused by the proposed merging rule consisting of the grey relation analysis of similarity and local area energy. Then, grey absolute relation analysis is again utilized as measurement indicator to estimate the similarities between the initial fused image and source images, during which the initial map is acquired and then corrected by the mathematical morphological opening and closing. Finally, the fused image is obtained with the guidance of the corrected map, namely the decision map. Experiment results demonstrate that the fused image using the proposed algorithm is more continuous in focused region and more similar to source images in brightness compared with state-of-art multi-focus image fusion algorithms, such as Curvelet transform, lifting stationary wavelet transform (LSWT), non-subsampled contourlet transform (NSCT) and non-subsampled shearlet transform (NSST). Meanwhile, the proposed method shows better superiority in term of the computational complexity.

Compression-Dependent Transform-Domain Downward Conversion for Block-Based Image Coding.

Transform-domain downward conversion (TDDC) for image coding is usually implemented by discarding some high-frequency components from each transformed block. As a result, a block of fewer coefficients is formed, and a lower compression cost is achieved due to the coding of only a few low-frequency coefficients. In this paper, we focus on the design of a new TDDC-based coding method by using our proposed interpolation-compression directed filtering (ICDF) and error-compensated scalar quantization (ECSQ), leading to the compression-dependent TDDC (CDTDDC)-based coding. More specifically, ICDF is first used to convert each macro-block into an coefficient block. Then, this coefficient block is compressed with ECSQ, resulting in a smaller compression distortion for those pixels that locate at some specific positions of a macro-block. We select these positions according to the 4:1 uniform sub-sampling lattice and use the pixels locating at them to reconstruct the whole macro-block through an interpolation. The proposed CDTDDC-based coding can be applied to compress both grayscale and color images. More importantly, when it is used in the color image compression, it offers not only a new solution to reduce the data-size of chrominance components but also a higher compression efficiency. Experimental results demonstrate that applying our proposed CDTDDC-based coding to compress still images can achieve a significant quality gain over the existing compression methods.

Change Detection Algorithm Based on Discrete Wavelet Transforms and Neighborhood Fuzzy C-Means

A new algorithm on Discrete Wavelet Transform (DWT) and neighborhood FCM is proposed to detect change area from remote sensing image. First, the subtraction and ratio image are obtained by the subtraction and ratio method from the two registered remote sensing images; Then, the DWT is applied to the subtraction and ratio image, the region intensity-based and energy-based fusion rules is adopted to the low frequency and high frequency wavelet coefficients, and the inverse DWT is used to obtain the final difference image; At last, the neighborhood FCM is carried out to get the change areas, the spatial distance information and gray difference information are considered in the objective function of FCM, which could avoid misclassification and enhance the detection probability. Experimental results show that the proposed algorithm has strong ability to suppress noise and good detection results; the detection probability of unban change area can reach to 98.45%, whereas, the detection probability is up to 87.5% for the discontinuous forest change area.

Frequency subspace amplitude flow for phase retrieval.

A novel approach, termed frequency subspace amplitude flow (FSAF), is proposed to reconstruct complex-valued signal from "phaseless" measurements. The proposed FSAF consists of two stages: the first stage approximates low-frequency coefficients of an unknown signal by the spectral method, and the second stage refines the results by the truncated conjugate gradient of amplitude-based nonconvex formulation. FSAF is easy to implement and applicable to natural images, where no additional constraint is needed. Extensive experiments with 1D signals, 2D images, and natural images corroborate significant improvements by using the proposed FSAF method over the state of the art. Especially for sample complexity, FSAF pushes the state of the art for exactly reconstructing complex natural signals (with a size of n) from 3.2n to 2.2n under the Gaussian model, and from 5n to 3n under the coherent diffraction pattern (CDP) model without increasing computational complexity. More importantly, the proposed method is highly flexible and can be easily adapted to the existing algorithms under different noise models.

Local energy match based non-sub sampled contourlet Transform for multi modal medical image fusion

Image processing technology requires moreover the full image or the part of image which is to be processed from the user’s point of view like the radius of object etc. The main purpose of fusion is to diminish dissimilar error between the fused image and the input images. With respect to the medical diagnosis, the edges and outlines of the concerned objects is more important than extra information. So preserving the edge features of the image is worth for investigating the image fusion. The image with higher contrast contains more edge-like features. Here we propose a new medical image fusion scheme namely Local Energy Match NSCT based on discrete contourlet transformation, which is constructive to give the details of curve edges. It is used to progress the edge information of fused image by dropping the distortion. This transformation lead to crumbling of multimodal image addicted to finer and coarser details and finest details will be decayed into unusual resolution in dissimilar orientation. The input multimodal images namely CT and MRI images are first transformed by Non Sub sampled Contourlet Transformation (NSCT) which decomposes the image into low frequency and high frequency elements. In our system, the Low frequency coefficient of the image is fused by image averaging and Gabor filter bank algorithm. The processed High frequency coefficients of the image are fused by image averaging and gradient based fusion algorithm. Then the fused image is obtained by inverse NSCT with local energy match based coefficients. To evaluate the image fusion accuracy, Peak Signal to Noise Ratio (PSNR), Root Mean Square Error (RMSE) and Correlation Coefficient parameters are used in this work .

Ferroelectret nanogenerator with large transverse piezoelectric activity

Energy harvesting from vibrations enables power delivery to low-energy electronics for autonomous devices, wearables, as well as wireless and remote sensing. Here we present a piezoelectric energy harvester consisting of specially designed ferroelectrets based on fluorocarbon polymers that have very large low-frequency transverse piezoelectric coefficients g31 = 3.0 Vm/N. This is much larger than corresponding values for lead zirconate titanate (PZT) or for polyvinylidene fluoride (PVDF) which are presently standard materials for harvester devices. We use these ferroelectrets in miniature energy harvesters that can be classified as Ferroelectret Nanogenerators (FENG's). Their performance is characterized by a power output of approximately 50 µW for an acceleration of 9.81 m/s2 and a seismic mass of 0.09 g. This compares favorably with the best PZT or PVDF systems, but is accomplished with a soft material that is small, lightweight, and flexible. This permits uses of such harvesters for powering miniature devices, such as small mobile sensing equipment, nano- and microelectronic circuits, body-worn electronics, or medical implants.

A blind and robust video watermarking based on IWT and new 3D generalized chaotic sine map

In the modern world, high-bandwidth internet access has led to the theft of all types of digital data, including digital video, and their illegal distribution. The digital video watermarking technique was introduced late last century to enforce video copyright protection. In this paper, a robust blind and secure video watermarking method is presented based on integer wavelet transform and the generalized chaotic sine map. In this method, integer wavelet transform is applied to each main frame of the standard video. Basically, watermarking techniques are evaluated based on three concepts of content quality, data resilience and data capacity. Therefore, in order to guarantee the quality of the watermarked video, the watermark is inserted in low-frequency coefficients. An appropriate security level is added to increase the efficiency and functionality by chaotic map to watermark. In addition, the normalized correlation coefficient (NCC) between the main watermark and extraction watermark is used as the main criterion of resistance measurement. The results show that the proposed method is good in terms of both quality and resistance to a variety of attacks, and the NCC values obtained in most cases are 1 or very close to 1.

The application of ECG cancellation in diaphragmatic electromyographic by using stationary wavelet transform.

In this paper, we present and investigate a special kind of stationary wavelet algorithm using "inverse" hard threshold to eliminate the electrocardiogram (ECG) interference included in diaphragmatic electromyographic (EMGdi). Differing from traditional wavelet hard threshold, "inverse" hard threshold is used to shrink strong coefficients of ECG interference and reserve weak coefficients of EMGdi signal. Meanwhile, a novel QRS location algorithm is proposed for the position detection of R wave by using low frequency coefficients in this paper. With the proposed method, raw EMGdi is decomposed by wavelet at fifth scale. Then, each ECG interference threshold is calculated by mean square, which is estimated by wavelet coefficients in the ECG cycle at each level. Finally, ECG interference wavelet coefficients are removed by "inverse" hard threshold, and then the de-noised signal is reconstructed by wavelet coefficients. The simulation and clinical EMGdi de-noising results show that the "inverse" hard threshold investigated in this paper removes the ECG interference in EMGdi availably and reserves its signal characteristics effectively, as compared to wavelet threshold.

Low frequency complex dielectric (conductivity) response of dilute clay suspensions: Modeling and experiments

In this work, we establish an effective medium model to describe the low-frequency complex dielectric (conductivity) dispersion of dilute clay suspensions. We use previously obtained low-frequency polarization coefficients for a charged oblate spheroidal particle immersed in an electrolyte as the building block for the Maxwell Garnett mixing formula to model the dilute clay suspension. The complex conductivity phase dispersion exhibits a near-resonance peak when the clay grains have a narrow size distribution. The peak frequency is associated with the size distribution as well as the shape of clay grains and is often referred to as the characteristic frequency. In contrast, if the size of the clay grains has a broad distribution, the phase peak is broadened and can disappear into the background of the canonical phase response of the brine. To benchmark our model, the low-frequency dispersion of the complex conductivity of dilute clay suspensions is measured using a four-point impedance measurement, which can be reliably calibrated in the frequency range between 0.1 Hz and 10 kHz. By using a minimal number of fitting parameters when reliable information is available as input for the model and carefully examining the issue of potential over-fitting, we found that our model can be used to fit the measured dispersion of the complex conductivity with reasonable parameters. The good match between the modeled and experimental complex conductivity dispersion allows us to argue that our simplified model captures the essential physics for describing the low-frequency dispersion of the complex conductivity of dilute clay suspensions.

DCT-Based Preprocessing Approach for ICA in Hyperspectral Data Analysis.

The huge quantity of information and the high spectral resolution of hyperspectral imagery present a challenge when performing traditional processing techniques such as classification. Dimensionality and noise reduction improves both efficiency and accuracy, while retaining essential information. Among the many dimensionality reduction methods, Independent Component Analysis (ICA) is one of the most popular techniques. However, ICA is computationally costly, and given the absence of specific criteria for component selection, constrains its application in high-dimension data analysis. To overcome this limitation, we propose a novel approach that applies Discrete Cosine Transform (DCT) as preprocessing for ICA. Our method exploits the unique capacity of DCT to pack signal energy in few low-frequency coefficients, thus reducing noise and computation time. Subsequently, ICA is applied on this reduced data to make the output components as independent as possible for subsequent hyperspectral classification. To evaluate this novel approach, the reduced data using (1) ICA without preprocessing; (2) ICA with the commonly used preprocessing techniques which is Principal Component Analysis (PCA); and (3) ICA with DCT preprocessing are tested with Support Vector Machine (SVM) and K-Nearest Neighbor (K-NN) classifiers on two real hyperspectral datasets. Experimental results in both instances indicate that data after our proposed DCT preprocessing method combined with ICA yields superior hyperspectral classification accuracy.

Comparative Study of Image Reconstruction Based on Compressive Sensing

Compressive sensing based image reconstruction that improves the algorithm to applying different approach which is DWT and DCT. First, by using wavelet transform, wavelet low frequency of the sub bands in which the image is decomposed in to low frequency and high frequency wavelet coefficients, second is to applied DCT on low frequency coordinates and construct the different transformation matrix. Use the measurement matrix measure the high frequency coefficient components and combine with DCT low frequency components image and sparse signal output is applied on compressive sensing. In compressive sensing, random measurement matrices are generally used and ?1minimisation algorithms often use linear programming to cover sparse signal vectors. But explicitly constructible measurement matrices providing performance guarantees were and ?1minimisation algorithms are often demanding in computational complexity for applications involving very large problem dimensions. To improve the PSNR (pick signal to noise ratio) of reconstructions image uses different coding such as Huffman and Arithmetic.

A new SAR image despeckling using correlation based fusion and method noise thresholding

This paper presents a new technique for despeckling of Synthetic Aperture Radar (SAR) images using a local correlation based fusion of high-frequency coefficients in Discrete Wavelet Transform (DWT) with method noise thresholding. The decomposition level is decided by analyzing the texture of the input image at each level by calculating entropy. The core idea of the proposed technique lies in the selection of decomposition level in 2D-DWT based on entropy parameter and on the fusion of high-frequency coefficients. On decomposition, the low-frequency coefficients remain untouched and the high-frequency coefficients are thresholded using two different shrinkage rules. Therefore the Bayesian and Bivariate shrinkage methods are applied to the high-frequency coefficients. After performing two different thresholding methods, the improved high-frequency coefficients are fused using local correlation based strategy. The threshold value is calculated by correlation strategy. Later the correlation coefficient (CC) is evaluated between the two improved high-frequency coefficients. The CC is now compared with the threshold value for the fusion purpose. On the basis of defined fusion strategy, the average and maximum operation are applied to perform the fusion of high-frequency coefficients. The despeckling scheme is followed by method noise thresholding in order to preserve the fine details of the image. The performance of the proposed method is assessed using metrics such as Signal-to-Noise Ratio (SNR), Peak-Signal-to-Noise Ratio (PSNR), Structural Similarity Index Metric (SSIM) and visual appearance of the despeckled image. The experimental results demonstrate the effectiveness of proposed work over prior works on SAR image despeckling.

Convolutional neural network-based multimodal image fusion via similarity learning in the shearlet domain

Recently, deep learning has been shown effectiveness in multimodal image fusion. In this paper, we propose a fusion method for CT and MR medical images based on convolutional neural network (CNN) in the shearlet domain. We initialize the Siamese fully convolutional neural network with a pre-trained architecture learned from natural data; then, we train it with medical images in a transfer learning fashion. Training dataset is made of positive and negative patch pair of shearlet coefficients. Examples are fed in two-stream deep CNN to extract features maps; then, a similarity metric learning based on cross-correlation is performed aiming to learn mapping between features. The minimization of the logistic loss objective function is applied with stochastic gradient descent. Consequently, the fusion process flow starts by decomposing source CT and MR images by the non-subsampled shearlet transform into several subimages. High-frequency subbands are fused based on weighted normalized cross-correlation between feature maps given by the extraction part of the CNN, while low-frequency coefficients are combined using local energy. Training and test datasets include pairs of pre-registered CT and MRI taken from the Harvard Medical School database. Visual analysis and objective assessment proved that the proposed deep architecture provides state-of-the-art performance in terms of subjective and objective assessment. The potential of the proposed CNN for multi-focus image fusion is exhibited in the experiments.

Dual image watermarking by exploiting the properties of selected DCT coefficients with JND modeling

The distinctive properties of partly sign-altered (PSA) coefficients in discrete cosine transform (DCT) domain are explored to achieve effective dual blind image watermarking. A host image is divided into non-overlapped blocks, each of which is then converted to a DCT representation separately. For each block, low-frequency DCT coefficients are selected for dual binary watermarking. The first binary bit is applied to the mean value of the PSA coefficients. Our formulation takes into account of human psychovisual characteristics. A stair-like quantization function is designed to not only regulate embedding strength but accommodate extra embedding for blocks with high contrast masking. In case second watermarking is requested, we tactically adjust the standard deviation of the PSA coefficients while keeping the PSA mean intact. The embedding strengths with respect to the involved parameters in the first and second watermarking are analytically examined. Experiment results indicate that the proposed scheme is capable of achieving excellent robustness at a peak signal-to-noise ratio (PSNR) around 38.7 dB. In particular, the exploitation of just noticeable distortion helps to enhance the robustness of the first watermark without causing noticeable quality degradation. The secondary watermark is embedded at the cost of approximately 1 dB PSNR. Yet the resultant performance is comparable with other compared methods in terms of bit error rates.

Robust Image Watermarking Algorithm Based on ASIFT against Geometric Attacks

Image processing technology has been developed rapidly in recent years, and altering the content of an image is easy for everyone, but may be illegal for an attacker. Thus, it is urgent and necessary to overcome this problem to protect the integrity and authenticity of images. Watermarking is a powerful technique proposed to solve this problem. This paper introduces a robust image watermarking algorithm working in the wavelet domain, embedding the watermark information into the third level low frequency coefficients after the three-level discrete wavelet transform (DWT) and singular value decomposition (SVD). Additionally, to improve the robustness to geometric attacks, the affine-scale-invariant feature transform (ASIFT) is applied to obtain feature points which are invariant to geometric attacks. Then, features of acquired points between the watermarked image and the received image are used to realize the resynchronization to improve the robustness. Experimental results show that the proposed algorithm achieves great balance between robustness and imperceptibility, and is robust against geometric attacks, JPEG compression, noise addition, cropping, median filters, and so on.

Microscopy mineral image enhancement based on improved adaptive threshold in nonsubsampled shearlet transform domain

In this paper, a novel microscopy mineral image enhancement method based on adaptive threshold in non-subsampled shearlet transform (NSST) domain is proposed. First, the image is decomposed into one low-frequency sub-band and several high-frequency sub-bands. Second, the gamma correction is applied to process the low-frequency sub-band coefficients, and the improved adaptive threshold is adopted to suppress the noise of the high-frequency sub-bands coefficients. Third, the processed coefficients are reconstructed with the inverse NSST. Finally, the unsharp filter is used to enhance the details of the reconstructed image. Experimental results on various microscopy mineral images demonstrated that the proposed approach has a better enhancement effect in terms of objective metric and subjective metric.