Bayesian Image Reconstruction Research Articles

A data-driven approach for the partial reconstruction of individual human molar teeth using generative deep learning.

Due to the high prevalence of dental caries, fixed dental restorations are regularly required to restore compromised teeth or replace missing teeth while retaining function and aesthetic appearance. The fabrication of dental restorations, however, remains challenging due to the complexity of the human masticatory system as well as the unique morphology of each individual dentition. Adaptation and reworking are frequently required during the insertion of fixed dental prostheses (FDPs), which increase cost and treatment time. This article proposes a data-driven approach for the partial reconstruction of occlusal surfaces based on a data set that comprises 92 3D mesh files of full dental crown restorations. A Generative Adversarial Network (GAN) is considered for the given task in view of its ability to represent extensive data sets in an unsupervised manner with a wide variety of applications. Having demonstrated good capabilities in terms of image quality and training stability, StyleGAN-2 has been chosen as the main network for generating the occlusal surfaces. A 2D projection method is proposed in order to generate 2D representations of the provided 3D tooth data set for integration with the StyleGAN architecture. The reconstruction capabilities of the trained network are demonstrated by means of 4 common inlay types using a Bayesian Image Reconstruction method. This involves pre-processing the data in order to extract the necessary information of the tooth preparations required for the used method as well as the modification of the initial reconstruction loss. The reconstruction process yields satisfactory visual and quantitative results for all preparations with a root mean square error (RMSE) ranging from 0.02 mm to 0.18 mm. When compared against a clinical procedure for CAD inlay fabrication, the group of dentists preferred the GAN-based restorations for 3 of the total 4 inlay geometries. This article shows the effectiveness of the StyleGAN architecture with a downstream optimization process for the reconstruction of 4 different inlay geometries. The independence of the reconstruction process and the initial training of the GAN enables the application of the method for arbitrary inlay geometries without time-consuming retraining of the GAN.

The first spatio-spectral Bayesian imaging of SN1006 in X-rays

Supernovae (SNs) are an important source of energy in the interstellar medium. Young remnants of supernovae (SNRs) exhibit peak emission in the X-ray region, making them interesting objects for X-ray observations. In particular, the supernova remnant SN1006 is of great interest due to its historical record, proximity, and brightness. Thus, it has been studied with a number of X-ray telescopes. Improving X-ray imaging of this and other remnants is an important but challenging task, as it often requires multiple observations with different instrument responses to image the entire object. Here, we use Chandra observations to demonstrate the capabilities of Bayesian image reconstruction using information field theory (IFT). Our objective is to reconstruct denoised, deconvolved, and spatio-spectral resolved images from X-ray observations and to decompose the emission into different morphologies, namely, diffuse and point-like. Further, we aim to fuse data from different detectors and pointings into a mosaic and quantify the uncertainty of our result. By utilizing prior knowledge on the spatial and spectral correlation structure of the diffuse emission and point sources, this method allows for the effective decomposition of the signal into these two components. In order to accelerate the imaging process, we introduced a multi-step approach, in which the spatial reconstruction obtained for a single energy range is used to derive an informed starting point for the full spatio-spectral reconstruction. We applied this method to 11 Chandra observations of SN1006 from 2008 and 2012, providing a detailed, denoised, and decomposed view of the remnant. In particular, the separated view of the diffuse emission ought to provide new insights into the complex, small-scale structures in the center of the remnant and at the shock front profiles. For example, our analysis reveals sharp X-ray flux increases by up to two orders of magnitude at the shock fronts of SN1006.

Contributed Session II: Computational modeling of shift in unique yellow for small stimuli.

Unique yellow (UY) is largely invariant to L:M cone proportion for spatially-extended stimuli in healthy trichromats. However, a recent adaptive-optics-based study by Boehm et al. reveals that when stimulus size is reduced to a few arcmin, color appearance depends on the local L:M proportion in the patch of the retina on which the stimulus was imaged. We aimed to determine if such findings are consistent with a normative account of visual processing. A series of 3.5 and 10 arcmin stimuli were simulated as isoluminant mixtures of 540 and 680 nm primaries. We modeled sensory encoding under adaptive-optics conditions using the open-source software ISETBio, for simulated retinal cone mosaics with varying local L:M proportions. The resultant cone excitations were decoded using a Bayesian image reconstruction algorithm (Zhang et al., 2022). For the 3.5 arcmin stimuli, as local L:M proportion decreased, the 540 nm component of the reconstructions increased relative to the 680 nm component. This is qualitatively consistent with the experimental observations of Boehm et al. For 10 arcmin stimuli, in contrast, reconstructions were stable across variation in local L:M cone proportion. Notably, reconstructions depend not only on the local L:M cone proportion, but also on the proportion in the immediately surrounding retina, leading to a testable prediction. The computational observations frame the experimental results as a normative consequence of visual processing.

Rapid single-photon color imaging of moving objects.

This paper outlines an experimental demonstration of a Bayesian image reconstruction approach to achieve rapid single-photon color imaging of moving objects. The capacity to extract the color of objects is important in a variety of target identification and computer vision applications. Nonetheless, it remains challenging to achieve high-speed color imaging of moving objects in low-photon flux environments. The low-photon regime presents particular challenges for efficient spectral separation and identification, while unsupervised image reconstruction algorithms are often slow and computationally expensive. In this paper, we address both of these difficulties using a combination of hardware and computational solutions. We demonstrate color imaging using a Single-Photon Avalanche Diode (SPAD) detector array for rapid, low-light-level data acquisition, with an integrated color filter array (CFA) for efficient spectral unmixing. High-speed image reconstruction is achieved using a bespoke Bayesian algorithm to produce high-fidelity color videos. The analysis is conducted first on simulated data allowing different pixel formats and photon flux scenarios to be investigated. Experiments are then performed using a plasmonic metasurface-based CFA, integrated with a 64 × 64 pixel format SPAD array. Passive imaging is conducted using white-light illumination of multi-colored, moving targets. Intensity information is recorded in a series of 2D photon-counting SPAD frames, from which accurate color information is extracted using the fast Bayesian method introduced herein. The per-frame reconstruction rate proves to be hundreds of times faster than the previous computational method. Furthermore, this approach yields additional information in the form of uncertainty measures, which can be used to assist with imaging system optimization and decision-making in real-world applications. The techniques demonstrated point the way towards rapid video-rate single-photon color imaging. The developed Bayesian algorithm, along with more advanced SPAD technology and utilization of time-correlated single-photon counting (TCSPC) will permit live 3D, color videography in extremely low-photon flux environments.

A Joint-Parameter Estimation and Bayesian Reconstruction Approach to Low-Dose CT.

Most penalized maximum likelihood methods for tomographic image reconstruction based on Bayes' law include a freely adjustable hyperparameter to balance the data fidelity term and the prior/penalty term for a specific noise-resolution tradeoff. The hyperparameter is determined empirically via a trial-and-error fashion in many applications, which then selects the optimal result from multiple iterative reconstructions. These penalized methods are not only time-consuming by their iterative nature, but also require manual adjustment. This study aims to investigate a theory-based strategy for Bayesian image reconstruction without a freely adjustable hyperparameter, to substantially save time and computational resources. The Bayesian image reconstruction problem is formulated by two probability density functions (PDFs), one for the data fidelity term and the other for the prior term. When formulating these PDFs, we introduce two parameters. While these two parameters ensure the PDFs completely describe the data and prior terms, they cannot be determined by the acquired data; thus, they are called complete but unobservable parameters. Estimating these two parameters becomes possible under the conditional expectation and maximization for the image reconstruction, given the acquired data and the PDFs. This leads to an iterative algorithm, which jointly estimates the two parameters and computes the to-be reconstructed image by maximizing a posteriori probability, denoted as joint-parameter-Bayes. In addition to the theoretical formulation, comprehensive simulation experiments are performed to analyze the stopping criterion of the iterative joint-parameter-Bayes method. Finally, given the data, an optimal reconstruction is obtained without any freely adjustable hyperparameter by satisfying the PDF condition for both the data likelihood and the prior probability, and by satisfying the stopping criterion. Moreover, the stability of joint-parameter-Bayes is investigated through factors such as initialization, the PDF specification, and renormalization in an iterative manner. Both phantom simulation and clinical patient data results show that joint-parameter-Bayes can provide comparable reconstructed image quality compared to the conventional methods, but with much less reconstruction time. To see the response of the algorithm to different types of noise, three common noise models are introduced to the simulation data, including white Gaussian noise to post-log sinogram data, Poisson-like signal-dependent noise to post-log sinogram data and Poisson noise to the pre-log transmission data. The experimental outcomes of the white Gaussian noise reveal that the two parameters estimated by the joint-parameter-Bayes method agree well with simulations. It is observed that the parameter introduced to satisfy the prior's PDF is more sensitive to stopping the iteration process for all three noise models. A stability investigation showed that the initial image by filtered back projection is very robust. Clinical patient data demonstrated the effectiveness of the proposed joint-parameter-Bayes and stopping criterion.

Bayesian Image Reconstruction Using Weighted Laplace Prior for Lung Respiratory Monitoring With Electrical Impedance Tomography

The poor spatial resolution of electrical impedance tomography (EIT) limits its use in medical devices because of its highly nonlinear and ill-posed nature. A novel hierarchical block sparse Bayesian learning (BSBL) method is designed for lung respiratory monitoring with EIT. It is its excellent modeling capability and noise robustness that allows BSBL to adaptively explore and exploit the internal conductivity distribution, e.g., block sparsity and intra-block correlation. First, the <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$K$ </tex-math></inline-formula> -nearest neighbor (KNN) strategy is employed to automatically group each block based on the clustering property and incorporated to constrain the intra-block correlation matrix based on the spatial distance. Second, a three-layer hierarchical BSBL model using weighted Laplace (WL) prior is considered to enhance the recovery performance. Finally, an efficient bound optimization (BO) method is performed for Bayesian inference, avoiding the tedious parameter adjustment. This leads to improvements in recovery performance, algorithm robustness, and computational efficiency. Moreover, root mean square error (RMSE), image correlation coefficient (ICC), and relative size coverage ratio (RCR) are applied for quantitative comparisons of image quality. Numerical simulations and in vivo lung respiratory experiments showed that the proposed KNN-BSBL-WL method outperforms existing referenced methods in terms of reconstruction accuracy and computational time. On average, KNN-BSBL-WL obtains the RMSE <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$= 0.07\,\,\pm \,\,0.02$ </tex-math></inline-formula> , ICC <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$= 0.96\,\,\pm \,\,0.01$ </tex-math></inline-formula> , and RCR <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$= 0.98\,\,\pm \,\,0.05$ </tex-math></inline-formula> , which are closest to the true values of the recorded snapshots obtained in the experiments. Meanwhile, the average time of KNN-BSBL-WL is 0.16 s, achieving an acceleration ratio close to <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\sim 3$ </tex-math></inline-formula> compared with the referenced KNN-BSBL method. Therefore, the proposed method is an efficient and robust means of imaging reconstruction, which further improves the feasibility of EIT in respiratory clinical applications.

An image reconstruction framework for characterizing initial visual encoding.

We developed an image-computable observer model of the initial visual encoding that operates on natural image input, based on the framework of Bayesian image reconstruction from the excitations of the retinal cone mosaic. Our model extends previous work on ideal observer analysis and evaluation of performance beyond psychophysical discrimination, takes into account the statistical regularities of the visual environment, and provides a unifying framework for answering a wide range of questions regarding the visual front end. Using the error in the reconstructions as a metric, we analyzed variations of the number of different photoreceptor types on human retina as an optimal design problem. In addition, the reconstructions allow both visualization and quantification of information loss due to physiological optics and cone mosaic sampling, and how these vary with eccentricity. Furthermore, in simulations of color deficiencies and interferometric experiments, we found that the reconstructed images provide a reasonable proxy for modeling subjects' percepts. Lastly, we used the reconstruction-based observer for the analysis of psychophysical threshold, and found notable interactions between spatial frequency and chromatic direction in the resulting spatial contrast sensitivity function. Our method is widely applicable to experiments and applications in which the initial visual encoding plays an important role.

Bayesian Image Reconstruction from Retinal Cone Signals

Bayesian Image Reconstruction from Retinal Cone Signals

Iterative versus non-iterative image reconstruction methods for sparse magnetic resonance imaging.

Magnetic resonance imaging (MRI) using under-sampled k-space data is a common method to shorten the imaging time. Iterative Bayesian algorithms are usually used for its image reconstruction. This paper compares an iterative Bayesian image reconstruction method that uses both spatial and temporal constraints and a non-iterative reconstruction algorithm that does not use temporal constraints. Three patient studies are performed. It is interesting to notice that the images reconstructed by the iterative Bayesian algorithm may introduce more bias than the non-iterative algorithm, even though the images provided by the iterative Bayesian algorithm look less noisy. The bias can be reduced by decreasing the influence of the temporal constraints.

Hybrid Very Long Baseline Interferometry Imaging and Modeling with themis

Abstract Generating images from very long baseline interferometric observations poses a difficult, and generally not unique, inversion problem. This problem is simplified by the introduction of constraints, some generic (e.g., positivity of the intensity) and others motivated by physical considerations (e.g., smoothness, instrument resolution). It is further complicated by the need to simultaneously address instrumental systematic uncertainties and sparse coverage in the u–v plane. We report a new Bayesian image reconstruction technique in the parameter estimation framework Themis that has been developed for the Event Horizon Telescope. This has two key features: first, the full Bayesian treatment of the image reconstruction makes it possible to generate a full posterior for the images, permitting a rigorous and quantitative investigation into the statistical significance of image features. Second, it is possible to seamlessly incorporate directly modeled features simultaneously with image reconstruction. We demonstrate this second capability by incorporating a narrow, slashed ring in reconstructions of simulated M87 data in an attempt to detect and characterize the photon ring. We show that it is possible to obtain high-fidelity photon ring sizes, enabling mass measurements with accuracies of 2%–5% that are essentially insensitive to astrophysical uncertainties, and creating opportunities for precision tests of general relativity.

Artifact-free whole-slide imaging with structured illumination microscopy and Bayesian image reconstruction.

BackgroundStructured illumination microscopy (SIM) is a method that can be used to image biological samples and can achieve both optical sectioning and super-resolution effects. Optimization of the imaging set-up and data-processing methods results in high-quality images without artifacts due to mosaicking or due to the use of SIM methods. Reconstruction methods based on Bayesian estimation can be used to produce images with a resolution beyond that dictated by the optical system.FindingsFive complete datasets are presented including large panoramic SIM images of human tissues in pathophysiological conditions. Cancers of the prostate, skin, ovary, and breast, as well as tuberculosis of the lung, were imaged using SIM. The samples are available commercially and are standard histological preparations stained with hematoxylin-eosin.ConclusionThe use of fluorescence microscopy is increasing in histopathology. There is a need for methods that reduce artifacts caused by the use of image-stitching methods or optical sectioning methods such as SIM. Stitched SIM images produce results that may be useful for intraoperative histology. Releasing high-quality, full-slide images and related data will aid researchers in furthering the field of fluorescent histopathology.

Physics-Based Regularizer for Joint Soft Segmentation and Reconstruction of Electron Microscopy Images of Polycrystalline Microstructures

For Bayesian image reconstruction applications in which the measured image follows from physical considerations, it is desirable to incorporate the corresponding physics into the prior model. In this paper, we use a phase-field as the physics-based prior model to implement the soft segmentation and reconstruction of noisy microstructural images of a polycrystalline, covalent material (SiC). The functional form of this prior is based on a coarse-grained Ginzburg–Landau free energy that embodies the underlying physics, and its phenomenological parameters are obtained from atomistic computer simulation. In particular, we compare an existing functional form developed by Fan and Chen for microstructural simulations with one developed here that is better suited to noise reduction in image reconstruction, and find that the latter form is indeed superior in this context. Numerical and experimental results demonstrate that the proposed method performs successful soft segmentation and reconstruction of microscopy images, even at very low signal levels. In addition, the superior performance of the proposed model for several case studies in comparison with state-of-the-art methods, such as BM3D and one using a MRF-based prior, is demonstrated.

Spatially adaptive Bayesian image reconstruction through locally-modulated Markov random field models

The use of Markov random field (MRF) models has proven to be a fruitful approach in a wide range of image processing applications. It allows local texture information to be incorporated in a systematic and unified way and allows statistical inference theory to be applied giving rise to novel output summaries and enhanced image interpretation. A great advantage of such low-level approaches is that they lead to flexible models, which can be applied to a wide range of imaging problems without the need for significant modification. This paper proposes and explores the use of conditional MRF models for situations where multiple images are to be processed simultaneously, or where only a single image is to be reconstructed and a sequential approach is taken. Although the coupling of image intensity values is a special case of our approach, the main extension over previous proposals is to allow the direct coupling of other properties, such as smoothness or texture. This is achieved using a local modulating function which adjusts the influence of global smoothing without the need for a fully inhomogeneous prior model. Several modulating functions are considered and a detailed simulation study, motivated by remote sensing applications in archaeological geophysics, of conditional reconstruction is presented. The results demonstrate that a substantial improvement in the quality of the image reconstruction, in terms of errors and residuals, can be achieved using this approach, especially at locations with rapid changes in the underlying intensity.

Auditory entrainment of motor responses in older adults with and without Parkinson’s disease: An MEG study

Medical therapies applied to Parkinson’s disease (PD) have advanced tremendously since the 1960’s based on advances in our understanding of the underlying neurophysiology. Behavioral therapies, such as rhythmic auditory stimulation (RAS), have been developed more recently and demonstrated efficacy. However, the neural mechanisms of RAS are only vaguely understood. In this study, we examined the neurophysiology of RAS using magnetoencephalography (MEG) in a sample of older adults with (21 people) and without PD (23 participants). All participants underwent high-density MEG during a beat-based cued tapping task with rhythmic and non-rhythmic patterns, and the resulting data were analyzed using a Bayesian image reconstruction method. Complex wavelet based time-frequency decomposition was used to compute inter-trial phase locking factor (PLF) to auditory stimuli for left and right signal space projection vectors. Tapping with a rhythm compared to a non-rhythmic sequence resulted in differential brain activity in each group: (i) a greater activation of temporal, motor and parietal areas was found in healthy adults; (ii) a greater reliance on parietal and frontal gyri was found in PD participants. During rhythmic tapping, older adults without PD had significantly stronger neural activity in bilateral frontal, supplementary and primary motor areas compared to those with PD. Conversely, older adults with PD exhibited significantly stronger activity in the bilateral parietal regions, as well as the rolandic operculum and bilateral supramarginal gyri, relative to their healthy peers. These data suggest that RAS mobilizes diverse oscillatory networks; Healthy controls may shift to frontal areas mobilization whereas PD patients rely on parietal areas to a greater extent, which may reflect frontal network dysfunction with compensation in PD, and could serve as specific regions of interest for further RAS studies.

Imaging tissues and cells beyond the diffraction limit with structured illumination microscopy and Bayesian image reconstruction

BackgroundStructured illumination microscopy (SIM) is a family of methods in optical fluorescence microscopy that can achieve both optical sectioning and super-resolution effects. SIM is a valuable method for high-resolution imaging of fixed cells or tissues labeled with conventional fluorophores, as well as for imaging the dynamics of live cells expressing fluorescent protein constructs. In SIM, one acquires a set of images with shifting illumination patterns. This set of images is subsequently treated with image analysis algorithms to produce an image with reduced out-of-focus light (optical sectioning) and/or with improved resolution (super-resolution).FindingsFive complete, freely available SIM datasets are presented including raw and analyzed data. We report methods for image acquisition and analysis using open-source software along with examples of the resulting images when processed with different methods. We processed the data using established optical sectioning SIM and super-resolution SIM methods and with newer Bayesian restoration approaches that we are developing.ConclusionsVarious methods for SIM data acquisition and processing are actively being developed, but complete raw data from SIM experiments are not typically published. Publically available, high-quality raw data with examples of processed results will aid researchers when developing new methods in SIM. Biologists will also find interest in the high-resolution images of animal tissues and cells we acquired. All of the data were processed with SIMToolbox, an open-source and freely available software solution for SIM.

A bayesian method for accelerated magnetic resonance elastography of the liver.

PurposeMagnetic resonance elastography is a noninvasive tool for quantifying soft tissue stiffness. Magnetic resonance elastography has been adopted as a clinical method for staging liver fibrosis. However, the application of liver magnetic resonance elastography requires multiple lengthy breath‐holds. We propose a new data acquisition and processing method to reduce magnetic resonance elastography scan time.MethodsA Bayesian image reconstruction method that uses transform sparsity and magnitude consistency across different phase offsets to recover images from highly undersampled data is proposed. The method is validated using retrospectively down‐sampled phantom data and prospectively down‐sampled in vivo data (N = 86).ResultsThe proposed technique allows accurate quantification of mean liver stiffness up to an acceleration factor of R = 6, enabling acquisition of a slice in 4.3 s. Bland‐Altman analysis indicates that the proposed technique (R = 6) has a bias of −0.04 kPa and limits of agreement of −0.36 to + 0.28 kPa when compared with traditional generalized autocalibrating partial parallel acquisition reconstruction (R = 1.4).ConclusionBy exploiting transform sparsity and magnitude consistency, accurate quantification of mean stiffness in the liver can be obtained at an acceleration rate of up to R = 6. This potentially enables the collection of three to four liver slices, as per clinical protocol, within a single breath‐hold. Magn Reson Med 80:1178–1188, 2018. © 2018 International Society for Magnetic Resonance in Medicine.

Accelerated Gibbs sampling of normal distributions using matrix splittings and polynomials

Standard Gibbs sampling applied to a multivariate normal distribution with a specified precision matrix is equivalent in fundamental ways to the Gauss–Seidel iterative solution of linear equations in the precision matrix. Specifically, the iteration operators, the conditions under which convergence occurs, and geometric convergence factors (and rates) are identical. These results hold for arbitrary matrix splittings from classical iterative methods in numerical linear algebra giving easy access to mature results in that field, including existing convergence results for antithetic-variable Gibbs sampling, REGS sampling, and generalizations. Hence, efficient deterministic stationary relaxation schemes lead to efficient generalizations of Gibbs sampling. The technique of polynomial acceleration that significantly improves the convergence rate of an iterative solver derived from a symmetric matrix splitting may be applied to accelerate the equivalent generalized Gibbs sampler. Identicality of error polynomials guarantees convergence of the inhomogeneous Markov chain, while equality of convergence factors ensures that the optimal solver leads to the optimal sampler. Numerical examples are presented, including a Chebyshev accelerated SSOR Gibbs sampler applied to a stylized demonstration of low-level Bayesian image reconstruction in a large 3-dimensional linear inverse problem.

Resting-State Neurophysiological Abnormalities in Posttraumatic Stress Disorder: A Magnetoencephalography Study.

Posttraumatic stress disorder (PTSD) is a debilitating psychiatric condition that is common in veterans returning from combat operations. While the symptoms of PTSD have been extensively characterized, the neural mechanisms that underlie PTSD are only vaguely understood. In this study, we examined the neurophysiology of PTSD using magnetoencephalography (MEG) in a sample of veterans with and without PTSD. Our primary hypothesis was that veterans with PTSD would exhibit aberrant activity across multiple brain networks, especially those involving medial temporal and frontal regions. To this end, we examined a total of 51 USA combat veterans with a battery of clinical interviews and tests. Thirty-one of the combat veterans met diagnostic criteria for PTSD and the remaining 20 did not have PTSD. All participants then underwent high-density MEG during an eyes-closed resting-state task, and the resulting data were analyzed using a Bayesian image reconstruction method. Our results indicated that veterans with PTSD had significantly stronger neural activity in prefrontal, sensorimotor and temporal areas compared to those without PTSD. Veterans with PTSD also exhibited significantly stronger activity in the bilateral amygdalae, parahippocampal and hippocampal regions. Conversely, healthy veterans had stronger neural activity in the bilateral occipital cortices relative to veterans with PTSD. In conclusion, these data suggest that veterans with PTSD exhibit aberrant neural activation in multiple cortical areas, as well as medial temporal structures implicated in affective processing.

Bayesian image reconstruction for positron emission tomography based on anisotropic median-diffusion

For improving the quality of positron emission tomography (PET) images, the partial differential equation median (PDEmedian) algorithm which incorporates an anisotropic diffusion (AD) filter into the median root prior (MRP) algorithm was proposed. However, due to the shortcomings of the AD filter, the PDEmedian algorithm is difficult to realize. This work aims to solve this problem by introducing a new diffusion model into the PDEmedian. The proposed algorithm shows its positive effects on image reconstruction and denoising. Experimental results present that the new algorithm can preserve sharp edges while reducing noise at the same time. Furthermore, in comparison to other similar reconstruction algorithms, the proposed method is less sensitive to the value of the gradient threshold and the adjustment of the diffusion number.

Multispectral Bayesian reconstruction technique for real-time two color fluorescence microscopy

We have developed a real-time imaging method for two-color widefield fluorescence microscopy using a combined approach that integrates multi-spectral imaging and Bayesian image reconstruction technique.