Simulated Image Data Research Articles

Satellite recognition and pose estimation using homeomorphic manifold analysis

We propose a novel monocular vision-based framework for both satellite recognition and pose estimation, using homeomorphic manifold analysis. We use a unified conceptual manifold to represent continuous pose variation of all satellites in the visual input space, learn nonlinear function mapping from conceptual manifold representation to visual inputs, and decompose discrete category variation in the mapping coefficient space. Experimental results on a simulated image data set show the effectiveness and robustness of our approach.

One-dimensional voting scheme for circle and arc detection.

Circle detection is an important issue that has not been perfectly solved in automated image analysis to date. It is traditionally carried out via pixel-based 3D voting algorithms, involving tremendous computation and requiring huge storage space with questionable accuracy. In this report, a novel edge-section-based 1D voting algorithm is developed in circle detection to improve the detection rate and precision. Based on experiments with simulated image data and a ground-tested standard dataset, the novel scheme significantly outperformed all previous state-of-the-art schemes in detection rate and precision, and was comparable to the state of the art in processing speed.

Bayesian perspective-plane (BPP) with maximum likelihood searching for visual localization

The proposed "Perspective-Plane" in this paper is similar to the well-known "Perspective-n-Point (PnP)" or "Perspective-n-Line (PnL)" problems in computer vision. However, it has broader applications and potentials, because planar scenes are more widely available than control points or lines in daily life. We address this problem in the Bayesian framework and propose the "Bayesian Perspective-Plane (BPP)" algorithm, which can deal with more generalized constraints rather than type-specific ones. The BPP algorithm consists of three steps: 1) plane normal computation by maximum likelihood searching from Bayesian formulation; 2) plane distance computation; and 3) visual localization. In the first step, computation of the plane normal is formulated within the Bayesian framework, and is solved by using the proposed Maximum Likelihood Searching Model (MLS-M). Two searching modes of 2D and 1D are discussed. MLS-M can incorporate generalized planar and out-of-plane deterministic constraints. With the computed normal, the plane distance is recovered from a reference length or distance. The positions of the object or the camera can be determined afterwards. Extensions of the proposed BPP algorithm to deal with un-calibrated images and for camera calibration are discussed. The BPP algorithm has been tested with both simulation and real image data. In the experiments, the algorithm was applied to recover planar structure and localize objects by using different types of constraints. The 2D and 1D searching modes were illustrated for plane normal computation. The results demonstrate that the algorithm is accurate and generalized for object localization. Extensions of the proposed model for camera calibration were also illustrated in the experiment. The potential of the proposed algorithm was further demonstrated to solve the classic Perspective-Three-Point (P3P) problem and classify the solutions in the experiment. The proposed BPP algorithm suggests a practical and effective approach for visual localization.

Test bed for real-time image acquisition and processing systems based on FlexRIO, CameraLink, and EPICS

Image diagnostics are becoming standard ones in nuclear fusion. At present, images are typically analyzed off-line. However, real-time processing is occasionally required (for instance, hot-spot detection or pattern recognition tasks), which will be the objective for the next generation of fusion devices. In this paper, a test bed for image generation, acquisition, and real-time processing is presented. The proposed solution is built using a Camera Link simulator, a Camera Link frame-grabber, a PXIe chassis, and offers software interface with EPICS. The Camera Link simulator (PCIe card PCIe8 DVa C-Link from Engineering Design Team) generates simulated image data (for example, from video-movies stored in fusion databases) using a Camera Link interface to mimic the frame sequences produced with diagnostic cameras. The Camera Link frame-grabber (FlexRIO Solution from National Instruments) includes a field programmable gate array (FPGA) for image acquisition using a Camera Link interface; the FPGA allows for the codification of ad-hoc image processing algorithms using LabVIEW/FPGA software. The frame grabber is integrated in a PXIe chassis with system architecture similar to that of the ITER Fast Controllers, and the frame grabber provides a software interface with EPICS to program all of its functionalities, capture the images, and perform the required image processing. The use of these four elements allows for the implementation of a test bed system that permits the development and validation of real-time image processing techniques in an architecture that is fully compatible with that of the ITER Fast Controllers. This paper provides a specific example of a pattern search in a movie, its real-time implementation, and a performance analysis of the entire platform.

The effects of endmember selection on modelling impervious surfaces using spectral mixture analysis: a case study in Sydney, Australia

In urban areas, spectral mixture analysis (SMA) is a common technique for deriving the fractions of land covers within a pixel and information on the distribution of impervious surfaces. This study examined how the selection of endmembers affected the quantification of impervious surfaces using TM and ASTER imagery. Multiple subsets of endmembers derived using (1) extreme pixels from a minimum noise fraction (MNF) transformation, and (2) a manual approach using a priori knowledge of the study area were analysed. Two data sets were used to assess accuracy: (1) simulated image data comprising unmixed and mixed pixels of 10 typical and spectrally different urban land covers, and (2) detailed data derived from high-resolution aerial photography. The dimensionality of the imagery limited the number of endmembers, and as a result, unmixed land covers were modelled using multiple endmembers and some cells had abundance values that summed to more than one or were negative. The land covers of red roofs and concrete were the largest contributors to the error in impervious surfaces. The Sequential Maximum Angle Convex Cone (SMACC) endmember model was also used to unmix the images; however, the larger number of endmembers did not resolve the use of multiple endmembers to model the unmixed land covers and the accuracy was similar to that using SMA. The relationship between the pervious fraction estimated using the vegetation endmember and the ground reference data was stronger than that for the impervious fraction, although the fraction was underestimated. The problems in modelling highly variable impervious surfaces with a limited number of endmembers suggest that in urban environments with substantial vegetation, modelling the vegetation component as the inverse of the impervious fraction may lead to improved results.

ParceLiNGAM: A Causal Ordering Method Robust Against Latent Confounders

We consider learning a causal ordering of variables in a linear nongaussian acyclic model called LiNGAM. Several methods have been shown to consistently estimate a causal ordering assuming that all the model assumptions are correct. But the estimation results could be distorted if some assumptions are violated. In this letter, we propose a new algorithm for learning causal orders that is robust against one typical violation of the model assumptions: latent confounders. The key idea is to detect latent confounders by testing independence between estimated external influences and find subsets (parcels) that include variables unaffected by latent confounders. We demonstrate the effectiveness of our method using artificial data and simulated brain imaging data.

Accurate 3D reconstruction by a new PDS-OSEM algorithm for HRRT

State-of-the-art high resolution research tomography (HRRT) provides high resolution PET images with full 3D human brain scanning. But, a short time frame in dynamic study causes many problems related to the low counts in the acquired data. The PDS-OSEM algorithm was proposed to reconstruct the HRRT image with a high signal-to-noise ratio that provides accurate information for dynamic data. The new algorithm was evaluated by simulated image, empirical phantoms, and real human brain data. Meanwhile, the time activity curve was adopted to validate a reconstructed performance of dynamic data between PDS-OSEM and OP-OSEM algorithms. According to simulated and empirical studies, the PDS-OSEM algorithm reconstructs images with higher quality, higher accuracy, less noise, and less average sum of square error than those of OP-OSEM. The presented algorithm is useful to provide quality images under the condition of low count rates in dynamic studies with a short scan time.

Robust Independent Component Analysis via Minimum $\gamma $-Divergence Estimation

Independent component analysis (ICA) has been shown to be useful in many applications. However, most ICA methods are sensitive to data contamination. In this article we introduce a general minimum U-divergence framework for ICA, which covers some standard ICA methods as special cases. Within the U-family we further focus on the γ-divergence due to its desirable property of super robustness for outliers, which gives the proposed method γ-ICA. Statistical properties and technical conditions for recovery consistency of γ-ICA are studied. In the limiting case, it improves the recovery condition of MLE-ICA known in the literature by giving necessary and sufficient condition. Since the parameter of interest in γ-ICA is an orthogonal matrix, a geometrical algorithm based on gradient flows on special orthogonal group is introduced. Furthermore, a data-driven selection for the γ value, which is critical to the achievement of γ-ICA, is developed. The performance, especially the robustness, of γ-ICA is demonstrated through experimental studies using simulated data and image data.

Detection of Brain Activity in Functional Magnetic Resonance Imaging Data using Matrix Factorization

Non-negative matrix factorization (NMF) is becoming a popular tool for decomposition of data in the field of signal and image processing like Independent Component Analysis (ICA). In this study we are relaxing the requirement of non-negative data for NMF making the update equations simple and thus making it Matrix Factorization (MF) and implementing it on simulated Functional Magnetic Resonance Imaging (fMRI) data for detection of neuronal activity. Simulated fMRI data is processed to detect the hidden sources of task related activity, functional activity and artifacts using the proposed MF technique. Performance of the proposed scheme is better than NMF in terms of average correlation results of the extracted sources/time courses with the actual sources/time courses. Similarly proposed MF is computationally cost effective and converges fast as compared to NMF. Also extracted sources obey no permutation which is the limitation of ICA and NMF.

A flexible and robust approach for segmenting cell nuclei from 2D microscopy images using supervised learning and template matching

We describe a new supervised learning-based template matching approach for segmenting cell nuclei from microscopy images. The method uses examples selected by a user for building a statistical model that captures the texture and shape variations of the nuclear structures from a given dataset to be segmented. Segmentation of subsequent, unlabeled, images is then performed by finding the model instance that best matches (in the normalized cross correlation sense) local neighborhood in the input image. We demonstrate the application of our method to segmenting nuclei from a variety of imaging modalities, and quantitatively compare our results to several other methods. Quantitative results using both simulated and real image data show that, while certain methods may work well for certain imaging modalities, our software is able to obtain high accuracy across several imaging modalities studied. Results also demonstrate that, relative to several existing methods, the template-based method we propose presents increased robustness in the sense of better handling variations in illumination, variations in texture from different imaging modalities, providing more smooth and accurate segmentation borders, as well as handling better cluttered nuclei.

A Kernel-Based Target-Constrained Interference-Minimized Filter for Hyperspectral Sub-Pixel Target Detection

The target-constrained interference-minimized filter (TCIMF) method has been successfully applied to various hyperspectral target detection applications. This paper presents a nonlinear version of TCIMF, called kernel-based TCIMF (KTCIMF), employing the kernel method to resolve the issue of nonlinear endmember mixing in hyperspectral images (HSI). Input data are implicitly mapped into a high-dimensional feature space, where it is assumed that target signals are more separable from background signals. Conventional TCIMF performs well in suppressing undesired signatures whose spectra are similar to that of the targets, thereby enhancing performance, and with less false alarms. KTCIMF not only takes into consideration the nonlinear endmember mixture but also fully exploits the other spectrally similar interference signatures. In this way, it is effective in suppressing both the background and those undesired signatures that may cause false alarms in traditional methods. Experimental results with both simulated and real hyperspectral image data confirm KTCIMF's performance with intimately mixed data. Compared with conventional kernel detectors, KTCIMF shows improved ROC curves and better separability between targets and backgrounds.

Generation of a suite of 3D computer‐generated breast phantoms from a limited set of human subject data

The authors previously reported on a three-dimensional computer-generated breast phantom, based on empirical human image data, including a realistic finite-element based compression model that was capable of simulating multimodality imaging data. The computerized breast phantoms are a hybrid of two phantom generation techniques, combining empirical breast CT (bCT) data with flexible computer graphics techniques. However, to date, these phantoms have been based on single human subjects. In this paper, the authors report on a new method to generate multiple phantoms, simulating additional subjects from the limited set of original dedicated breast CT data. The authors developed an image morphing technique to construct new phantoms by gradually transitioning between two human subject datasets, with the potential to generate hundreds of additional pseudoindependent phantoms from the limited bCT cases. The authors conducted a preliminary subjective assessment with a limited number of observers (n = 4) to illustrate how realistic the simulated images generated with the pseudoindependent phantoms appeared. Several mesh-based geometric transformations were developed to generate distorted breast datasets from the original human subject data. Segmented bCT data from two different human subjects were used as the "base" and "target" for morphing. Several combinations of transformations were applied to morph between the "base' and "target" datasets such as changing the breast shape, rotating the glandular data, and changing the distribution of the glandular tissue. Following the morphing, regions of skin and fat were assigned to the morphed dataset in order to appropriately assign mechanical properties during the compression simulation. The resulting morphed breast was compressed using a finite element algorithm and simulated mammograms were generated using techniques described previously. Sixty-two simulated mammograms, generated from morphing three human subject datasets, were used in a preliminary observer evaluation where four board certified breast radiologists with varying amounts of experience ranked the level of realism (from 1 = "fake" to 10 = "real") of the simulated images. The morphing technique was able to successfully generate new and unique morphed datasets from the original human subject data. The radiologists evaluated the realism of simulated mammograms generated from the morphed and unmorphed human subject datasets and scored the realism with an average ranking of 5.87 ± 1.99, confirming that overall the phantom image datasets appeared more "real" than "fake." Moreover, there was not a significant difference (p > 0.1) between the realism of the unmorphed datasets (6.0 ± 1.95) compared to the morphed datasets (5.86 ± 1.99). Three of the four observers had overall average rankings of 6.89 ± 0.89, 6.9 ± 1.24, 6.76 ± 1.22, whereas the fourth observer ranked them noticeably lower at 2.94 ± 0.7. This work presents a technique that can be used to generate a suite of realistic computerized breast phantoms from a limited number of human subjects. This suite of flexible breast phantoms can be used for multimodality imaging research to provide a known truth while concurrently producing realistic simulated imaging data.

Magneto-Optic Imaging for Aircraft Skins Inspection: A Probability of Detection Study of Simulated and Experimental Image Data

The increasing fleet of aging aircrafts has resulted in an increasing demand for cost effective nondestructive evaluation (NDE) techniques that are accurate, reliable, and easy to use. Magneto-Optic Imaging (MOI) is such a technique, which has gained wide acceptance for detection of both surface and subsurface defects in multi-layer aircraft structures. The main advantage of MOI is rapid inspection and ease of interpreting image data in contrast to complex impedance signals from conventional eddy current instruments. One missing piece of the puzzle for advanced MOI systems is how to quantitatively analyse the MO images, and understand the detectability limits when image data are acquired under varying operational conditions. This paper presents a probability of detection (POD) study that is conducted using both simulation model-predicted and experimental MO image data. Simulated panels from a 3-D FEM model and experimental panels with machined defects are used to generate data for interpretation by human inspectors or automated systems, and subsequently for POD studies. The POD curves demonstrate the merits in optimizing inspection parameters that maximized the performance of current MOI systems. Parameters quantifying the detectability of MO image data using skewness functions are also presented and discussed.

Combining Hopfield Neural Network and Contouring Methods to Enhance Super-Resolution Mapping

The mixed pixel problem may be reduced through the use of a soft image classification and super-resolution mapping analyses. Here, the positive attributes of two popular super-resolution mapping methods, based on contouring and the Hopfield neural network, are combined. For a binary classification scenario, the method is based on fitting a contour of equal class membership to a pre-final output of a standard Hopfield neural network. Analyses of simulated and real image data sets show that the proposed method is more accurate than the standard contouring and Hopfield neural network based methods, with error typically reduced by a factor of two or more. The sensitivity of the Hopfield neural network based approaches to the setting of a gain function is also explored.

Stochastic 3D modeling of La0.6Sr0.4CoO3−δ cathodes based on structural segmentation of FIB–SEM images

A stochastic microstructure model is developed in order to describe and simulate the 3D geometry of two-phase microstructures (solid and pore phase), where the solid phase consists of spherical particles being completely connected with each other. Such materials appear e.g. in La0.6Sr0.4CoO3−δ (LSC) cathodes of solid oxide fuel cells, which are produced by screen printing and sintering of a paste consisting of LSC powder manufactured by flame spray synthesis. Thus, as a model type, we consider (fully parameterized) random sphere systems which are based on ideas from stochastic geometry and graph theory. In particular, the midpoints of spheres are modeled by random point processes. In order to assure the complete connectivity of the spheres, a modified version of the relative neighborhood graph is introduced. This graph controls the radii of spheres such that a completely connected sphere system is obtained. The model parameters are exemplarily fitted to three different materials for LSC cathodes, produced with sintering temperatures of 750, 850 and 950°C, respectively. Finally, the goodness of fit is validated by comparing structural characteristics of real and simulated image data.

Accuracy and Reliability of Automated Gray Matter Segmentation Pathways on Real and Simulated Structural Magnetic Resonance Images of the Human Brain

Automated gray matter segmentation of magnetic resonance imaging data is essential for morphometric analyses of the brain, particularly when large sample sizes are investigated. However, although detection of small structural brain differences may fundamentally depend on the method used, both accuracy and reliability of different automated segmentation algorithms have rarely been compared. Here, performance of the segmentation algorithms provided by SPM8, VBM8, FSL and FreeSurfer was quantified on simulated and real magnetic resonance imaging data. First, accuracy was assessed by comparing segmentations of twenty simulated and 18 real T1 images with corresponding ground truth images. Second, reliability was determined in ten T1 images from the same subject and in ten T1 images of different subjects scanned twice. Third, the impact of preprocessing steps on segmentation accuracy was investigated. VBM8 showed a very high accuracy and a very high reliability. FSL achieved the highest accuracy but demonstrated poor reliability and FreeSurfer showed the lowest accuracy, but high reliability. An universally valid recommendation on how to implement morphometric analyses is not warranted due to the vast number of scanning and analysis parameters. However, our analysis suggests that researchers can optimize their individual processing procedures with respect to final segmentation quality and exemplifies adequate performance criteria.

An investigation into factors affecting electron density calibration for a megavoltage cone‐beam CT system

There is a growing interest in the use of megavoltage cone‐beam computed tomography (MV CBCT) data for radiotherapy treatment planning. To calculate accurate dose distributions, knowledge of the electron density (ED) of the tissues being irradiated is required. In the case of MV CBCT, it is necessary to determine a calibration‐relating CT number to ED, utilizing the photon beam produced for MV CBCT. A number of different parameters can affect this calibration. This study was undertaken on the Siemens MV CBCT system, MVision, to evaluate the effect of the following parameters on the reconstructed CT pixel value to ED calibration: the number of monitor units (MUs) used (5, 8, 15 and 60 MUs), the image reconstruction filter (head and neck, and pelvis), reconstruction matrix size (256 by 256 and 512 by 512), and the addition of extra solid water surrounding the ED phantom. A Gammex electron density CT phantom containing EDs from 0.292 to 1.707 was imaged under each of these conditions. The linear relationship between MV CBCT pixel value and ED was demonstrated for all MU settings and over the range of EDs. Changes in MU number did not dramatically alter the MV CBCT ED calibration. The use of different reconstruction filters was found to affect the MV CBCT ED calibration, as was the addition of solid water surrounding the phantom. Dose distributions from treatment plans calculated with simulated image data from a 15 MU head and neck reconstruction filter MV CBCT image and a MV CBCT ED calibration curve from the image data parameters and a 15 MU pelvis reconstruction filter showed small and clinically insignificant differences. Thus, the use of a single MV CBCT ED calibration curve is unlikely to result in any clinical differences. However, to ensure minimal uncertainties in dose reporting, MV CBCT ED calibration measurements could be carried out using parameter‐specific calibration measurements.PACS number: 87.59.bd

초분광 영상자료의 Endmember 추출 속도 향상에 관한 연구

본 연구에서는 MNF 변환 후 각 밴드별 최소, 최대값을 가지는 화소들을 추출하고, 이 후보 화소들만을 대상으로 endmemeber를 추출하는 기법을 제안하고 이의 적용 가능성을 고찰하였다. 이와 같은 접근은 영상자료의 크기와 관계없이 밴드수<TEX>${\times}$</TEX>2개에 해당하는 화소들만을 대상으로 endmember를 추출하므로 전체 화소를 대상으로 endmember 여부를 계산하는 기존의 기법들보다 계산시간 측면에서 매우 효율적일 수 있다. 모의 초분광 영상자료 및 미국 네바다 Cuprite 지역의 AVIRIS 초분광 영상자료에 N-FINDR 기법을 적용하여 endmember를 추출함에 있어 본 기법을 적용해본 결과, 전체 화소를 대상으로 처리한 결과와 거의 동일한 결과를 얻을 수 있음을 확인하였다. 잡음에 의한 영항과 후보 화소 추출기준 등에 대한 추가 연구를 통해 대용량 초분광 영상자료의 처리 및 분석에 효율적으로 활용될 수 있을 것으로 기대된다. A fast algorithm for endmember extraction is proposed in this study which extracts min. and max. pixels from each band after MNF transform as candidate pixels for endmember. This method finds endmembers not from the entire image pixels but only from the previously extracted candidate pixels. The experimental results by N-FINDR using a simulated hyperspectral image data and AVIRIS Cuprite image data showed that the proposed fast algorithm extracts the same endmembers with the conventional methods. More studies on the effect of noise and more adaptive criteria in extracting candidate pixels are expected to increase the usability of this method for more fast and efficient analysis of hyperspectral image data.

Myocardial Blood Flow Quantification From MRI by Deconvolution Using an Exponential Approximation Basis

We have evaluated the use of deconvolution using an exponential approximation basis for the quantification of myocardial blood flow from perfusion cardiovascular magnetic resonance. Our experiments, based on simulated signal intensity curves, phantom acquisitions, and clinical image data, indicate that exponential deconvolution allows for accurate quantification of myocardial blood flow. Together with automated respiratory motion correction myocardial contour delineation, the exponential deconvolution enables efficient and reproducible quantification of myocardial blood flow in clinical routine.

Polynomial fitting of DT-MRI fiber tracts allows accurate estimation of muscle architectural parameters

Fiber curvature is a functionally significant muscle structural property, but its estimation from diffusion-tensor magnetic resonance imaging fiber tracking data may be confounded by noise. The purpose of this study was to investigate the use of polynomial fitting of fiber tracts for improving the accuracy and precision of fiber curvature (κ) measurements. Simulated image data sets were created in order to provide data with known values for κ and pennation angle (θ). Simulations were designed to test the effects of increasing inherent fiber curvature (3.8, 7.9, 11.8 and 15.3 m−1), signal-to-noise ratio (50, 75, 100 and 150) and voxel geometry (13.8- and 27.0-mm3 voxel volume with isotropic resolution; 13.5-mm3 volume with an aspect ratio of 4.0) on κ and θ measurements. In the originally reconstructed tracts, θ was estimated accurately under most curvature and all imaging conditions studied; however, the estimates of κ were imprecise and inaccurate. Fitting the tracts to second-order polynomial functions provided accurate and precise estimates of κ for all conditions except very high curvature (κ=15.3 m−1), while preserving the accuracy of the θ estimates. Similarly, polynomial fitting of in vivo fiber tracking data reduced the κ values of fitted tracts from those of unfitted tracts and did not change the θ values. Polynomial fitting of fiber tracts allows accurate estimation of physiologically reasonable values of κ, while preserving the accuracy of θ estimation.