Hidden Markov Tree Research Articles

EGGDD: An explicit dependency model for multi-modal medical image fusion in shift-invariant shearlet transform domain

Most of the traditional medical image fusion methods that use the multi-scale decomposition schemes suffer from the bad image representations and the loss of the dependency in different highpass subbands. To deal with these problems, a novel dependency model, named Explicit Generalized Gaussian Density Dependency (EGGDD) model, is developed by the shift-invariant shearlet transform (SIST). Substantially different from describing the dependency by two hidden states in the Hidden Markov Tree (HMT) model, we provide the scheme to explicitly describe the marginal statistics of each highpass subband using the Generalized Gaussian Density (GGD), as well as the dependency between different subbands by the Kullback–Leibler distance (KLD). After embedding the obtained dependency into each highpass subband, an efficient fusion scheme, inspired by the divisive normalization response in the V1 visual cortex model, is developed to combine the highpass-subband coefficients. The experiments demonstrate that the developed method can produce better fusion results than those of some existing methods by the comparison of visual sense and quantitative measurements in terms of mutual information, entropy, spatial frequency, standard deviation, QAB/F and QW.

Image Fusion Method Based on Universal Hidden Markov Tree Model in the Contourlet Domain

Considering the statistical characteristics of contourlet coefficients of images, we propose an image fusion method based on the universal contourlet hidden Markov tree (uHMT) model. A salience measure and a match measure are presented according to the probability of a contourlet coefficient belonging to the high state of the uHMT model which needs no training. Experimental results prove the effectiveness of our method in the field of visual quality and objective evaluations.

LiDAR simulation in modelled orchards to optimise the use of terrestrial laser scanners and derived vegetative measures

Light detection and ranging (LiDAR) technology is beginning to have an impact on agriculture. Canopy volume and/or fruit tree leaf area can be estimated using terrestrial laser sensors based on this technology. However, the use of these devices may have different options depending on the resolution and scanning mode. As a consequence, data accuracy and LiDAR derived parameters are affected by sensor configuration, and may vary according to vegetative characteristics of tree crops. Given this scenario, users and suppliers of these devices need to know how to use the sensor in each case. This paper presents a computer program to determine the best configuration, allowing simulation and evaluation of different LiDAR configurations in various tree structures (or training systems). The ultimate goal is to optimise the use of laser scanners in field operations. The software presented generates a virtual orchard, and then allows the scanning simulation with a laser sensor. Trees are created using a hidden Markov tree (HMT) model. Varying the foliar structure of the orchard the LiDAR simulation was applied to twenty different artificially created orchards with or without leaves from two positions (lateral and zenith). To validate the laser sensor configuration, leaf surface of simulated trees was compared with the parameters obtained by LiDAR measurements: the impacted leaf area, the impacted total area (leaves and wood), and the impacted area in the three outer layers of leaves.

Exact Inference of Hidden Markov Tree for Image Segmentation

Image segmentation methods that exploit multiscale information about images to be estimated have been extensively studied, typically using the Hidden Markov Tree (HMT) framework. we incorporate wavelet coefficients information of the original image in the form of Hidden Markov Tree model prior for the object segmentation. In this paper, we derive a generalized closed form inference scheme to exact determine the posterior likelihood at each iteration with definite number of iteration steps. Extensive experiments show that this method performs better than many competitive multiscale image segmentation methods.

Functional MR image statistical restoration for neural activity detection using hidden Markov tree model

In this paper, we present a framework for functional MR image restoration based on the Hidden Markov Tree (HMT) model. Under this scheme, the wavelet/contourlet coefficients of the distorted image are filtered using the HMT model of the baseline image to minimise the statistical divergence between two images. An iterative algorithm between image registration and HMT filtering is developed to achieve a trade-off between the least mean square error (in the spatial domain) and the minimum statistical divergence (in the spectral domain). We demonstrate that the proposed method can eliminate the motion artefacts (such as spikes and burring) in the Functional MR Imaging data more effectively, leading to reliable neural activity detection. This method can also be used for image restoration in other medical imaging applications.

Distributed Compressed Video Sensing in Camera Sensor Networks

With the booming of video devices ranging from low-power visual sensors to mobile phones, the video sequences captured by these simple devices must be compressed easily and reconstructed by relatively more powerful servers. In such scenarios, distributed compressed video sensing (DCVS), combining distributed video coding (DVC) and compressed sensing (CS), is developed as a novel and powerful signal-sensing and compression algorithm for video signals. In DCVS, video frames can be compressed to a few measurements in a separate manner, while the interframe correlation is explored by the joint recovery algorithm. In this paper, a new DCVS joint recovery scheme using side-information-based belief propagation (SI-BP) is proposed to exploit both the intraframe and interframe correlations, which is particularly efficient over error-prone channels. The DCVS scheme using SI-BP is designed over two frame signal models, the mixture Gaussian (MG) model and the wavelet hidden Markov tree (WHMT) model. Simulation results evaluated on two video sequences illustrate that the SI-BP-based DCVS scheme is error resilient when the measurements are transmitted through the noisy wireless channels.

Accelerated 3D MERGE carotid imaging using compressed sensing with a hidden markov tree model

To determine the potential for accelerated 3D carotid magnetic resonance imaging (MRI) using wavelet based compressed sensing (CS) with a hidden Markov tree (HMT) model. We retrospectively applied HMT model-based CS and conventional CS to 3D carotid MRI data with 0.7 mm isotropic resolution from six subjects with known carotid stenosis (12 carotids). We applied a wavelet-tree model learned from a training database of carotid images to improve CS reconstruction. Quantitative endpoints such as lumen area, wall area, mean and maximum wall thickness, plaque calcification, and necrotic core area were measured and compared using Bland-Altman analysis along with image quality. Rate-4.5 acceleration with HMT model-based CS provided image quality comparable to that of rate-3 acceleration with conventional CS and fully sampled reference reconstructions. Morphological measurements made on rate-4.5 HMT model-based CS reconstructions were in good agreement with measurements made on fully sampled reference images. There was no significant bias or correlation between mean and difference of measurements when comparing rate 4.5 HMT model-based CS with fully sampled reference images. HMT model-based CS can potentially be used to accelerate clinical carotid MRI by a factor of 4.5 without impacting diagnostic quality or quantitative endpoints.

Improved statistical image fusion method using a continuous-valued blur factor

Considering the reality of image formation as well as the statistical characteristics of contourlet coefficients, we propose a method for statistical image fusion in the contourlet domain. For high-frequency subbands, an image formation model was established that has a continuous-valued blur factor to reflect the actual imaging situation. According to this model, fused contourlet coefficients were estimated by use of an expectation-maximization (EM) algorithm. During the estimation, a contourlet hidden Markov tree model was adopted to grasp all dependencies among coefficients and aim for better estimation results. Because the blur factor is a continuous variable, we exploited an explicit expression to update the factor in the EM algorithm, which contributed to a decline in the number of iterations for convergence. Experimental results indicated that, especially for multifocus images, the proposed method provides more satisfying fusion results in terms of visual effects and objective evaluations than some typical fusion methods based on multiscale decomposition and some statistical fusion methods using a discrete blur factor.

Multi-modal medical image fusion using the inter-scale and intra-scale dependencies between image shift-invariant shearlet coefficients

For the quality of the fused outcome is determined by the amount of the information captured from the source images, thus, a multi-modal medical image fusion method is developed in the shift-invariant shearlet transform (SIST) domain. The two-state Hidden Markov Tree (HMT) model is extended into the SIST domain to describe the dependent relationships of the SIST coefficients of the cross-scale and inter-subbands. Base on the model, we explain why the conventional Average–Maximum fusion scheme is not the best rule for medical image fusion, and therefore a new scheme is developed, where the probability density function and standard deviation of the SIST coefficients are employed to calculate the fused coefficients. Finally, the fused image is obtained by directly applying the inverse SIST. Integrating the SIST and the HMT model, more spatial feature information of the singularities and more functional information contents can be preserved and transferred into the fused results. Visual and statistical analyses demonstrate that the fusion quality can be significantly improved over that of five typical methods in terms of entropy and mutual information, edgeinformation, standarddeviation, peak signal to noise and structural similarity. Besides, color distortion can be suppressed to a great extent, providing a better visual sense.

Oil Spill in SAR Image Denoising Method Based on Contourlet HMT

Contourlet domain hidden markov trees model was introduced in the SAR image denosing research work. Compared to the classical denosing method, the denoised SAR image can preserve information of oil spill and advantages for following up work. According to the gray level, this method segmented the SAR image to region of oil spill and ocean. Then using the method of Contourlet domain hidden markov trees model denoised the different regions. Experimental results show that the method represents better performance in speckle reduction and edges information detection.

InSAR Interferogram Filtering Using Wavelet-Domain Hidden Markov Tree Model

Speckle noise is an intrinsic property of InSAR interferograms,and InSAR interferograms are sensitive to it.In order to obtain a more accurate interferometric product,a noise filtering step must be performed.According to the statistical property of speckle noise,SAR interferograms filtering using wavelet-domain hidden markov tree(HMT) model is studied.It is employed to filter the real component and the imaginary component of interferogram respectively.The result shows that the new method is powerful to interferogram speckle noise suppression and residues reduction,as well as it can be preserved fine phase details and fringes' structures.The method proposed can improve the accuracy of interferometric products.

Temporal Feature Characterization via Dynamic Hidden Markov Tree

We present a novel multiscale dynamic methodology for automatic machine vision inspection aiming at characterizing temporal features of tobacco leaves. The image sequences of tobacco leaves are transformed from RGB color space to L*a*b* color space, which provides a uniform perceptual difference measure. The image sequences are then represented by a multiscale Dynamic Hidden Markov tree (DHMT), which models not only inter and intra scale dependences of wavelet coefficients, but also temporal dependences of foreground/background observational properties. Experimental results demonstrate temporal consistent mean and covariance values of model coefficients in a* color channel.

Compressive Imaging Using Approximate Message Passing and a Markov-Tree Prior

We propose a novel algorithm for compressive imaging that exploits both the sparsity and persistence across scales found in the 2D wavelet transform coefficients of natural images. Like other recent works, we model wavelet structure using a hidden Markov tree (HMT) but, unlike other works, ours is based on loopy belief propagation (LBP). For LBP, we adopt a recently proposed “turbo” message passing schedule that alternates between exploitation of HMT structure and exploitation of compressive-measurement structure. For the latter, we leverage Donoho, Maleki, and Montanari's recently proposed approximate message passing (AMP) algorithm. Experiments with a large image database suggest that, relative to existing schemes, our turbo LBP approach yields state-of-the-art reconstruction performance with substantial reduction in complexity.

Empirical Study of Wavelet Domain Image Joint Statistics and Proposition of an Efficient Correlation Map

This paper presents an empirical study of joint wavelet statistics for textures and other imagery to find an efficient correlation neighborhood. Since there is an established realization that modeling wavelet and other x-let coefficient relationships is crucial to any successful transform domain algorithm (such as Hidden Markov Trees), new works have been devoted to examine these dependencies from different aspects and propose an appropriate model. Because the time and computation complexity involved both in analyzing non-linear dependencies and in solving dependent models may restrict us to consider only a very small subset of contributing neighbors we focus our attention on linear dependencies (correlations) while having a squint on non-linear relations too. In this process, we study a collection of 5000 real images to corroborate our statistical analysis of the joint coefficient behavior and try to find an efficient and at the same time frugal relation map through different statistical means. The statistical observations are then certified by a coefficient significance measure and the competitiveness of the map is substantiated by plugging it into two dependent denoising frameworks.

A static image coding algorithm based on contourlet Classified Hidden Markov Tree model

Purpose – Because of advantages such as multi‐scale and multi‐direction, contourlet transform is better at treating a 2‐D image than wavelet transform and the Classified Hidden Markov Tree (CHMT) model is able to analyze statistical character of coefficients, which means they can be analyzed more efficiently and effectively. So in this paper, the purpose is to use contourlet transform and CHMT model to redevelop traditional set partitioning in hierarchical trees (SPIHT) code algorithm, and by using these two methods, hope to decrease the distortion and increase the quality.Design/methodology/approach – In this paper, the algorithm is divided into two parts: code part and decode part. As all processes are operated in the contourlet domain, contourlet transform is finished at the beginning of the code part. SPIHT algorithm in the contourlet domain will code these contourlet coefficients. CHMT model will be built in decode part, it will optimize decoded coefficients by calculating their statistical relations...

Detection of Transient Signals in Lung Sounds: Local Approach Using a Markovian Tree with Frequency Selectivity

We deal in this paper with the extraction of multiresolution statistical signatures for the characterization of transient signals in strongly noisy contexts. These short-time signals have sharp and highly variable frequency components. The time/frequency window to choose for our analysis is then a major issue. We have chosen the Wavelet Packet Transform (WPT) due to its ability to provide multiple windows analysis with different time/frequency resolutions. We propose a new oriented Hidden Markov Tree (HMT) dedicated to the tree structure of the WPT, which offers promising statistical characterization of time/frequency variations in a signal, by exploiting several time/frequency resolutions. This model is exploited in a Bayesian context for the segmentation of signals containing transient components. The chosen data likelihood is a Generalized Gaussian Distributions (GGD), well suited for the modeling of Wavelet Packet Coefficients (WPC) distributions. We demonstrate the efficiency of our method on synthetic signals with several Signal to Noise Ratio (SNR). Our application domain is related to biomedical signals, and more specifically for the detection of uprising abnormalities in pulmonary sounds. This original method shows remarkable ability to detect such sounds, which are usually buried in the normal lung noise and are often very difficult to perceive with the human earing.

Image Segmentation Based on Contextual Label Tree for Retrieval

Image segmentation is an important constituent portion in image processing and retrieval. Based on the traditional Wavelet-domain Hidden Markov Tree (HMT) Multi-scale Segmentation method, this paper presents a Contextual Label Tree (CLT) method according to the dependency information between image blocks belong to different scales, including the relation from the father node, the neighbor nodes and the neighbor nodes of the father. This method calculates the maximal similarity using context vectors that exit on every tree node and realizes image segmentation from coarse-scale to fine-scale. Experiments show that this method is satisfied with its segmentation performance.

SECOND GENERATION BANDELET-DOMAIN HIDDEN MARKOV TREE-3S MODEL FOR SAR IMAGE SEGMENTATION

Since the segmentation results of SAR images by traditional transform domain hidden Markov tree (HMT) model were unsatisfactory in homogenous regions and exact edges,a new segmentation method based on second generation bandelet-domain HMT-3S model was proposed.The method was called BHMT-3Sseg shortly.HMT-3S is a special kind of HMT which combines the correlation of different subbands.It is more reasonable to characterize texture regions than HMT model.BHMT-3Sseg modeled the second generation Bandelet coefficients of an image by using HMT-3S,and the SAR image segmentation results were obtained by training the parameters of HMT-3S and computing the likelihood of each scale and multiscale fusion based on a contextual model.The segmentation results by BHMT-3Sseg not only have more exact and more continuous edges,but also retain better region information.The experiments show that BHMT-3Sseg is efficient and effective for SAR image segmentation.

Contourlet-based super resolution restoration for remote sensing images

This paper presented a Contourlet-based super resolution for remote sensing images,which adopted Contourlet coefficients as the features.It described a better degree of directionality and anisotropy,and used the smallest Euclidean distance as the computed feature by global searching.According to the distributions of the found coefficients in finer scale,the Hidden Markov Tree(HMT)model was introduced to the remote sensing images in Contourlet domain.And the Expectation Maximization(EM)algorithm was applied to estimate the parameters of the HMT model.With the parameters,the Contourlet coefficients were renewed by using Bayesian estimation theory.Finally,the super resolution restorationfor remote sensing images has achieved better effect.

Super Resolution for Remote Sensing Images Based on a Universal Hidden Markov Tree Model

In this paper, we propose a new super resolution (SR) method called the maximum a posteriori based on a universal Hidden Markov Tree (HMT) model for remote sensing images. The HMT theory is first used to set up a prior model for reconstructing super resolved images from a sequence of warped, blurred, subsampled, and noise-contaminated low-resolution (LR) images. Because the wavelet coefficients of images can be well characterized as a mixed Gaussian distribution, an HMT model is better able to capture the dependences between multiscale wavelet coefficients. The new method is tested first against simulated LR views from a single Landsat7 panchromatic scene and, then, with actual data from four Landsat7 panchromatic images captured on different dates. Both tests show that our method achieves better SR results both visually and quantitatively than other methods.