Detail Subbands Research Articles

Infrared and visible image fusion based on oversampled graph filter banks

The infrared image (RI) and visible image (VI) fusion method merges complementary information from the infrared and visible imaging sensors to provide an effective way for understanding the scene. The graph filter bank-based graph wavelet transform possesses the advantages of the classic wavelet filter bank and graph representation of a signal. Therefore, we propose an RI and VI fusion method based on oversampled graph filter banks. Specifically, we consider the source images as signals on the regular graph and decompose them into the multiscale representations with M-channel oversampled graph filter banks. Then, the fusion rule for the low-frequency subband is constructed using the modified local coefficient of variation and the bilateral filter. The fusion maps of detail subbands are formed using the standard deviation-based local properties. Finally, the fusion image is obtained by applying the inverse transform on the fusion subband coefficients. The experimental results on benchmark images show the potential of the proposed method in the image fusion applications.

Frame-synchronous Blind Audio Watermarking for Tamper Proofing and Self-Recovery

This paper presents a lifting wavelet transform (LWT)-based blind audio watermarking scheme designed for tampering detection and self-recovery. Following 3-level LWT decomposition of a host audio, the coefficients in selected subbands are first partitioned into frames for watermarking. To suit different purposes of the watermarking applications, binary information is packed into two groups: frame-related data are embedded in the approximation subband using rational dither modulation; the source-channel coded bit sequence of the host audio is hidden inside the 2nd and 3rd -detail subbands using 2N-ary adaptive quantization index modulation. The frame-related data consists of a synchronization code used for frame alignment and a composite message gathered from four adjacent frames for content authentication. To endow the proposed watermarking scheme with a self-recovering capability, we resort to hashing comparison to identify tampered frames and adopt a Reed–Solomon code to correct symbol errors. The experiment results indicate that the proposed watermarking scheme can accurately locate and recover the tampered regions of the audio signal. The incorporation of the frame synchronization mechanism enables the proposed scheme to resist against cropping and replacement attacks, all of which were unsolvable by previous watermarking schemes. Furthermore, as revealed by the perceptual evaluation of audio quality measures, the quality degradation caused by watermark embedding is merely minor. With all the aforementioned merits, the proposed scheme can find various applications for ownership protection and content authentication.

A novel multi-focus image fusion method using multiscale shearing non-local guided averaging filter

The multiscale transform based image fusion method cannot effectively preserve detail information and easily produce artifacts. Faced with these problems, we present a novel multi-focus fusion method based on multiscale shearing non-local guided averaging filter (MSNLGA). First, we construct a new multiscale geometrical analysis tool called MSNLGA, which combines the non-local guided averaging filter with the shearing filter bank. The MSNLGA can represent the intrinsic geometric structure of image sparsely because of its good property in multiscale, multi-direction and shift-invariance. Then, the MSNLGA is used to decompose source images to obtain approximate subbands and directional detail subbands. For the approximate subbands, we extract the anti-noise spatial frequency feature from the source images to guide its fusion. For the directional detail subbands, we introduce the convolutional sparse representation, which is a model that can achieve sparse representation of an entire subband, to represent each subband so as to obtain the activity level measurement to fuse directional detail subbands. Finally, the fused image can be obtained by the inverse MSNLGA of the fused subbands. The experimental results show that the proposed method can be competitive with or even outperform the state-of-the-art fusion methods in terms of both visual and quantitative evaluations.

A code division multiplexing and block classification-based real-time reversible data-hiding algorithm for medical images

This paper proposes a novel reversible data hiding (RDH) scheme for medical image based on block classification and code division multiplexing (CDM) in frequency domain. In this scheme, the original medical image is divided into non-overlapping blocks and classified into texture and smooth groups by calculating the mean square error of each block. The texture blocks are then transformed into the frequency domain with integer-to-integer discrete wavelet transform method, and the secret data are embedded into the detail sub-bands of the chosen texture blocks with CDM-based RDH algorithm. Consequently, the detail information of medical image is clarified and the data-embedding capacity is enlarged. Simultaneously, as the spreading sequences employed for data embedding are orthogonal to each other, when the secret message is repeatedly embedded into the cover image, most of the elements of spreading sequences are mutually canceled, which enables the proposed scheme to obtain high visual quality even at large data-embedding capacities. Moreover, on the receiver side, only the receiver with the correct data-embedding and block classification parameters can reconstruct the secret message and the cover image completely; thus, the security of RDH is guaranteed. Experimental results have demonstrated that the proposed scheme can yield better overall performance than other state-of-the-art RDH schemes on medical images.

A Temporal Neural Trace of Wavelet Coefficients in Human Object Vision: An MEG Study.

Wavelet transform has been widely used in image and signal processing applications such as denoising and compression. In this study, we explore the relation of the wavelet representation of stimuli with MEG signals acquired from a human object recognition experiment. To investigate the signature of wavelet descriptors in the visual system, we apply five levels of multi-resolution wavelet decomposition to the stimuli presented to participants during MEG recording and extract the approximation and detail sub-bands (horizontal, vertical, diagonal) coefficients in each level of decomposition. Apart from, employing multivariate pattern analysis (MVPA), a linear support vector classifier (SVM) is trained and tested over the time on MEG pattern vectors to decode neural information. Then, we calculate the representational dissimilarity matrix (RDM) on each time point of the MEG data and also on wavelet descriptors using classifier accuracy and one minus Pearson correlation coefficient, respectively. Given the time-courses calculated from performing the Pearson correlation between the wavelet descriptors RDMs and MEG decoding accuracy in each time point, our result shows that the peak latency of the wavelet approximation time courses occurs later for higher level coefficients. Furthermore, studying the neural trace of detail sub-bands indicates that the overall number of statistically significant time points for the horizontal and vertical detail coefficients is noticeably higher than diagonal detail coefficients, confirming the evidence of the oblique effect that the horizontal and vertical lines are more decodable in the human brain.

Medical image fusion based on statistical modeling

For improving the imaging quality and increasing the clinical applicability, a novel approach to multimodal medical image fusion is proposed based on statistical modeling in contourlet transform domain. Firstly, the coefficients of the approximate subband are modeled as Gaussian mixture distribution, and fused by a new rule using the weighted average of a posterior probability. Then, the coefficients of the detail subbands are modeled by generalized Gaussian distribution, and a selection rule is used for fusion based on the estimated parameters and the matching measure. Finally, an effectively fused image is achieved through inverse contourlet transform. The experimental results show that, compared with existing approaches, the proposed method can make the fused image have better performance, and provide more valuable diagnostic information.

Frame-synchronized blind speech watermarking via improved adaptive mean modulation and perceptual-based additive modulation in DWT domain

This paper presents a blind speech watermarking algorithm that adopts different strategies to embed synchronization codes and information bits into separate DWT subbands. Speech frames for data hiding are chosen based on intensity thresholding. Under the guidance of the source-filter theory, a bipolar synchronization code sequence is substituted for the noisy part of the inverse filtered excitation in the 2nd level detail subband. Such a design allows the increase of embedding strength without adversely disturbing the pitch harmonics of voiced speech. Experiment results show that a synchronization code of size 640 is sufficient to render reliable detection. The PESQ metrics indicate that the quality degradation due to the synchronization code embedding is almost negligible. As for the binary embedding in the 2nd level approximation subband, we improve the adaptive mean modulation scheme to secure the retrieval of watermark bits on a frame basis. Experiment results confirm that, with the payload capacity set at 200 bit per second (bps), the proposed scheme demonstrates better robustness than three other DWT-based methods in the presence of commonly encountered signal processing attacks. Furthermore, reducing the capacity down to 100 bps alters the low-frequency spectral distribution, which contributes noteworthy improvements in robustness and imperceptibility.

Dynamic texture classification using Gumbel mixtures in the complex wavelet domain

Dynamic texture (DT) classification has attracted extensive attention in the field of image sequence analysis. The probability distribution model, which has been used to analysis DT, can describe well the distribution property of signals. Here, the authors introduce the finite mixtures of Gumbel distributions (MoGD) and the corresponding parameter estimation method based on expectation-maximisation algorithm. Then, the authors propose the DT features based on MoGD model for DT classification. Specifically, after decomposing DTs with the dual-tree complex wavelet transform (DT-CWT), the median values of complex wavelet coefficient magnitudes of non-overlapping blocks in detail subbands are modelled with MoGDs. The model parameters are accumulated into a feature vector to describe DT. During the classification, a variational approximation version of the Kullback-Leibler divergence is used to measure the similarity between different DTs. The experimental evaluations on two popular benchmark DT data sets (UCLA and DynTex++) demonstrate the effectiveness of the proposed approach.

A novel approach for salient object detection using double-density dual-tree complex wavelet transform in conjunction with superpixel segmentation

Salient object detection in wavelet domain has recently begun to attract researchers’ effort due to its desired ability to provide multi-scale analysis of an image simultaneously in both frequency and spatial domains. The proposed algorithm exploits the inherent multi-scale structure of the double-density dual-tree complex-oriented wavelet transform (DDDTCWT) to decompose each input image into four approximate sub-band images and 32 high-pass detailed sub-band images at each scale. These 32 detailed high-pass sub-bands at each scale are adequate to represent singularities of any geometric object with high precision and to mimic zooming-in and zooming-out process of human vision system. In the proposed model, we first compute a rough segmented saliency map (RSSM) by fusing multi-scale edge-to-texture features generated from DDDTCWT with segmentation results obtained from bipartite graph partitioning-based segmentation approach. Then, each pixel in RSSM is categorized into either background region or salient region based on a threshold. Finally, the pixels of the two regions are considered as samples to be drawn from a multivariate kernel function whose parameters are estimated using expectation maximization algorithm, to generate a saliency map. The performance of the proposed model is evaluated in terms of precision, recall, F-measure, area under the ROC curve and computation time using six publicly available image datasets. Extensive experimental results on six benchmark datasets demonstrate that the proposed model outperformed the existing 29 state-of-the-art methods in terms of F-measure on all five datasets, recall on four datasets and area under ROC curve on two datasets. In terms of mean recall value, mean F-measure value and mean AUC value on all six datasets, the proposed method outperforms all state-of-the-art methods. The proposed method also takes comparatively less computation time in comparison with many existing spatial domain methods.

A multiresolution processing method for contrast enhancement in portal imaging

Portal images have a unique feature among the imaging modalities used in radiotherapy: they provide direct visualization of the irradiated volumes. However, contrast and spatial resolution are strongly limited due to the high energy of the radiation sources. Because of this, imaging modalities using x-ray energy beams have gained importance in the verification of patient positioning, replacing portal imaging.The purpose of this work was to develop a method for the enhancement of local contrast in portal images. The method operates in the sub-bands of a wavelet decomposition of the image, re-scaling them in such a way that coefficients in the high and medium resolution sub-bands are amplified—an approach totally different from those operating on the image histogram, widely used nowadays.Portal images of an anthropomorphic phantom were acquired in an electronic portal imaging device. Then, different re-scaling strategies were investigated, studying the effects of the scaling parameters on the enhanced images. Also, the effect of using different types of transform was studied. Finally, the implemented methods were combined with histogram equalization methods like the contrast limited adaptive histogram equalization, and these combinations were compared.Uniform amplification of the detail sub-bands shows the best results in contrast enhancement. On the other hand, linear re-escalation of the high resolution sub-bands increases the visibility of fine detail in the images, at the expense of an increase in noise levels. Also, since processing is applied only to detail sub-bands, not to the approximation, the mean gray level of the image is minimally modified and no further display adjustments are required.It is shown that re-escalation of the detail sub-bands of portal images can be used as an efficient method for the enhancement of both the local contrast and the resolution of these images.

Iterative adaptive Despeckling SAR image using anisotropic diffusion filter and Bayesian estimation denoising in wavelet domain

In this paper, a new iterative algorithm has been presented by aggregating Stationary Wavelet Transform (SWT), Bilateral filtering, Bayesian estimation, and Anisotropic Diffusion (AD) filtering to reduce the speckle noise in SAR images. For this purpose, speckle images were first decomposed using two-dimensional stationary wavelet transform and then a suitable filtering method was used to filter respective coefficients of each sub-band of the speckled images. Generally, in wavelet transform-based noise reduction methods, filtering and thresholding techniques are usually applied to the coefficients of the detail sub-bands and the residual speckle noise is ignored in the approximate sub-band. In this paper, bilateral filtering has been applied to reduce the speckle noise in the approximate sub-band. We used Bayesian estimator to calculate the noise-free signal in the horizontal and vertical sub-bands with respect to that some parts of signal coefficients are eliminated in the traditional thresholding techniques. Moreover, we applied anisotropic diffusion filtering method to preserve the edges and structure of image along the diagonal subband which has more details (the entropy is maximum) than other directions in radar and optic images. Finally, both the proposed algorithm and other speckle noise reduction methods were applied on two synthetic speckled images and an actual SAR image in San Francisco. Their efficiencies were compared according to the Structural SIMilarity(SSIM), Peak Signal to Noise Ratio (PSNR), Equivalent Number of Looks (ENL), Speckle Suppression Index (SSI) and Speckle Suppression and Mean Preservation Index (SMPI). The experimental results indicate that the proposed algorithm efficiently reduces the speckle noise and preserves the edges and structure of image.

Hybrid Speckle Noise Reduction Method for Abdominal Circumference Segmentation of Fetal Ultrasound Images

Fetal biometric size such as abdominal circumference (AC) is used to predict fetal weight or gestational age in ultrasound images. The automatic biometric measurement can improve efficiency in the ultrasonography examination workflow. The unclear boundaries of the abdomen image and the speckle noise presence are the challenges for the automated AC measurement techniques. The main problem to improve the accuracy of the automatic AC segmentation is how to remove noise while retaining the boundary features of objects. In this paper, we proposed a hybrid ultrasound image denoising framework which was a combination of spatial-based filtering method and multiresolution based method. In this technique, an ultrasound image was decomposed into subbands using wavelet transform. A thresholding technique and the anisotropic diffusion method were applied to the detail subbands, at the same time the bilateral filtering modified the approximation subband. The proposed denoising approach had the best performance in the edge preservation level and could improve the accuracy of the abdominal circumference segmentation.

Multimodal medical image fusion using non-subsampled shearlet transform and pulse coupled neural network incorporated with morphological gradient

This research proposes a novel fusion scheme for non-subsampled shearlet transform (NSST) which is based on simplified model of pulse coupled neural network (PCNN). The images to be fused are acquired from Postgraduate Institute of Medical Education and Research, Chandigarh, India, and internet repository. The image database contains computed tomography and T2-weighted magnetic resonance images. The images to be fused are decomposed into approximation and detail sub-bands using NSST. The regional energy-based activity measure with consistency verification is applied to fuse the approximation sub-band of NSST. The novel morphological gradient of detail sub-bands is fed as external stimulus to PCNN to fuse detail sub-bands. The proposed method is compared with five state-of-the-art fusion schemes visually and using five fusion performance parameters. It is observed that the resultant images of the proposed fusion scheme show appropriate fusion characteristics and retain the bone, CSF and edema details in the clinical format required for disease evaluation by the radiologists. The proposed scheme requires lesser computational time than other state-of-the-art PCNN-based fusion schemes.

The multiscale directional neighborhood filter and its application to clutter removal in GPR data

We present a novel neighborhood filter (NF)-based clutter removal algorithm in ground-penetrating radar (GPR) images. Since NF uses only range kernel of the well-known bilateral filter, it is less complex and makes clutter removal method appropriate for real-time implementations. We extend NF to multiscale–multidirectional case: MDNF and then decompose the GPR image into approximation and detail subbands to capture the intrinsic geometrical structures that contain both target and clutter information. After directional decomposition, the clutter is eliminated by keeping the diagonal information for target component. Finally, the inverse transform is applied to the remaining subbands for reconstruction of clutter-free GPR image. Results of both simulated and real datasets validate the superiority of MDNF over the state-of-the-art methods, and it improves in the false alarm rate further by 5.5% at maximum detection performance.

Unsupervised change detection of flood affected areas in SAR images using Rayleigh‐based Bayesian thresholding

In the Bayesian context, the rationale for using a particular probability density function (pdf) for a synthetic aperture radar (SAR) image improves the performance of unsupervised change detection (CD). So, this work is concentrated on promoting the consequence of selecting an empirical pdf and its parameter estimation technique. Rayleigh pdf is used to model approximation and detail sub-bands of a dual tree complex wavelet transform (DTCWT) of the log ratio difference image. The chosen pdf has a potential impact on impulsive SAR images. The scale parameter of Rayleigh pdf is estimated using first-order moments by exploiting Mellin transform. This iteration-free parameter estimation technique, made the method faster. In each sub-band at each level of DTCWT decomposition, the binary threshold is estimated to represent the change signal at different scales. The performance in the final change map is improved by combining intra-scale along with inter-scale sub-band coefficients. The proposed method is tested on three flooding SAR image data sets acquired from European Space Agency environmental satellite, European remote-sensing satellite (ERS)-2 and ERS-1 SAR instruments. The experimental results show that the proposed method achieves good CD performance in terms of false alarm rate and overall detection accuracy in less computing time.

Context-based watermarking cum chaotic encryption for medical images in telemedicine applications

In this paper, we propose a security system for secure transmission of medical images in telemedicine applications. The system couples an IWT-LSB watermarking and an encryption based on random permutation and chaos, to ensure confidentiality, integrity, authentication and nonrepudiation of medical images. We use IWT due to the sensitive nature of medical images and the need to retain diagnostic quality after image reconstruction. During the watermarking phase, the medical image is decomposed into wavelet sub-bands. Electronic patient record and extracted context information are then embedded in the least significant bits of the detail sub-band (host) coefficients. During encryption, the reconstructed watermarked medical image is randomly permuted and the permuted pixels diffused with a chaotic key stream to produce the cipher watermarked image. Experimental results and analyzes show that the system provides sufficient security against various forms of attacks. Furthermore, we propose a security architecture for the system.

Fusing DTCWT and LBP Based Features for Rotation, Illumination and Scale Invariant Texture Classification

Classification of texture images with different orientation, illumination, and scale changes is a challenging problem in computer vision and pattern recognition. This paper proposes two descriptors and uses them jointly to fulfill such task. One can obtain an image pyramid by applying dual-tree complex wavelet transform (DTCWT) on the original image, and generate local binary patterns (LBP) in DTCWT domain, called LBPDTCWT, as local texture features. Moreover, log-polar (LP) transform is applied on the original image, and the energies of DTCWT coefficients on detail subbands of the LP image, called LPDTCWTE, are taken as global texture features. We fuse the two kinds of features for texture classification, and the experimental results on benchmark data sets show that our proposed method can achieve better performance than other the state-of-the-art methods.

An efficient face recognition method using contourlet and curvelet transform

In the paper we propose a novel method for face recognition using contourlet transform (CNT) and curvelet transform (CLT) which improves rate of face recognition under different challenges. We obtain smooth contour information along different directions by applying CNT on the face image while CLT having multiscale, multidirectional and anisotropic properties has been employed to represent the edges more prominently. Pre-processed training images are decomposed up to fourth level using CNT and coefficients of directional subbands are analysed to obtain the features from the images. In another approach CLT has been applied on the pre-processed face images and considering scale of four and angle eight, different statistical features are extracted from the detail subbands. Finally, we integrate the features obtained from two approaches. High dimensionality of feature space has been reduced by selecting important features depending on the entropy of the transform coefficients. Selected features are applied to recognize the face images using support vector machine (SVM) classifier. Experimental results show that the proposed feature extraction method improves recognition accuracy compare to other methods and efficiently handle the effect of Gaussian noise as tested on JAFFE, ORL and FERET database.

A Combined Detail Enhancing Algorithm and Texture Feature Extraction Method for Supervised Classification of Remote Sensing Images

In this paper, we propose a supervised classification in multispectral satellite images based on a novel detail enhancing texture feature extraction algorithm. The multispectral training and test images are first given for pre-processing, which is decomposed into low-pass approximation and high-pass multi-directional subbands by wavelet based contourlet transform. High pass subbands are easily interfered by noise. Based upon the Normal Shrink technique, thresholding is applied in high frequency images to eliminate the noise. The intra and inter scale fusion rule is used to combine the approximation and detail subbands to form the enhanced image. The co-occurrence features are calculated by forming the gray level co-occurrence matrix on training and test images. Mahalanobis distance classifier is applied on the training and test data sets for effective classification. The experiment result shows that the overall accuracy is improved to 2.2% for (Test-1) 2% for (Test-2) and 3.2 5% for (Test-3) and kappa coefficient is improved to 0.02 for (Test-1) image 0.03 for (Test-2) image and 0.03 (Test-3) image.

A Novel Scene-Based Video Watermarking Scheme for Copyright Protection

Abstract Many illegal copies of original digital videos are being made, as they can be replicated perfectly through the Internet. Thus, it is extremely necessary to protect the copyrights of the owner and prevent illegal copying. This paper presents a novel approach to digital video watermarking for copyright protection using two different algorithms, whereby successive estimation of a statistical measure was used to detect scene boundaries and watermark was embedded in the detected scenes with discrete wavelet transform. Haar wavelet was used for decomposition. For embedding, the approaches used were (i) the detailed subband (LH subband) and (ii) the approximate subband (LL subband) of the cover video. Imperceptibility, robustness, and channel capacity were measured using both algorithms. The system was tested for robustness in the presence of 15 different attacks of five different categories, and, under multiple attacks, ensured that a wide spectrum of attack analysis has been done. The performance metrics measured included mean square error, peak signal-to-noise ratio, structural similarity index, normalized correlation, and bit error rate. The experimental results demonstrated the better visual imperceptibility and improved performance in terms of normalized correlation and bit error rate with embedding using the LL subband. Comparative analysis with existing schemes proved the improved robustness, better imperceptibility, and reduced computational time of both the proposed schemes.