Images derived from a “virtual phantom” can be useful in characterizing the performance of imaging systems. This has driven the development of virtual breast phantoms implemented in simulation environments. In breast imaging, several such phantoms have been proposed. We analyze the non-Gaussian statistical properties from three classes of virtual breast phantoms and compare them to similar statistics from a database of breast images. These include clustered-blob lumpy backgrounds (CBLBs), truncated binary textures, and the UPenn virtual breast phantoms. We use Laplacian fractional entropy (LFE) as a measure of the non-Gaussian statistical properties of each simulation procedure. Our results show that, despite similar power spectra, the simulation approaches differ considerably in LFE with very low scores for the CBLB to high values for the UPenn phantom at certain frequencies. These results suggest that LFE may have value in developing and tuning virtual phantom simulation procedures.
Virtual Clinical Trials (VCTs) of breast imaging have been used as a tool for the evaluation and optimization of novel imaging systems through computer simulations of breast anatomy, image acquisition, and interpretation. VCTs offer significant advantages over clinical trials in terms of cost, duration, and radiation risk. The performance of VCTs depends on the selection of simulated breasts to represent the population of interest. We have developed a method for selecting populations of software breast phantoms to match the clinical distribution of compressed breast thickness and breast percent density. We extracted the compressed thickness information from anonymized DICOM headers of mammography images from 10,705 women who had their breast screening exams within a year (09/2010-08/2011). Percent density was estimated using an open source software tool. Characteristic clinical sub-populations were identified by performing k-means clustering, and represented by separate sets of phantoms. The corresponding thickness of uncompressed phantoms was selected assuming 50% thickness reduction during mammographic compression. The phantom volumetric density was selected based upon a relationship between mammographic (2D) percent density and volumetric (3D) density, estimated from clinical images. Using a set of 24 representative phantoms, we were able to match the analyzed clinical population completely for the compressed breast thickness, and within two percentage points of the volumetric breast density. Representative phantoms can be used to generate the full population of virtual patients, of a size determined by the power-analysis of the specific VCT, by random variations of the internal phantom composition.
Virtual clinical trials (VCTs), computer simulations of clinical trials, can take many forms. In the field of breast imaging, VCTs often involve simulations of breast anatomy, which are used to produce simulated images of the breast with or without lesions. Our breast anatomy model consists of an array of voxels labeled to denote specific tissue types; the voxel labels are arrayed spatially so as to simulation various anatomic structures. Our most recent breast model includes numerous innovations in the anatomy simulation and data representation. The breast model has been revised in size and shape to better reflect the range of women seen clinically; the breast is divided into three breast regions (subcutaneous, interior, and posterior) with different rules to guide tissue arrangement; and tissue microstructure has been added to reflect a hierarchy of Cooper’s ligaments. The lesion simulation has been enhanced to support lesions with various shapes (e.g., spherical lesions with tapered periphery, circumscribed non-spherical lesions, and single or clustered microcalcifications) and lesion placement that follows the clinical prevalence. Finally, the data representation has been formalized to support large VCTs using the VCT pipeline software previously developed in our lab. These innovations have resulted in breast phantoms that are more realistic and more widely applicable.
A simulation of sequential breast pathology images is proposed, as a prerequisite for the development of virtual clinical trials (VCTs) with radiology-pathology (rad-path) correlation. The rad-path correlation of breast cancer findings is performed clinically to confirm concordance and increase confidence in diagnoses. VCTs have been used for optimization of breast imaging systems, based upon computer simulation of breast anatomy, imaging modalities, and image interpretation. Today, VCTs are used to optimize breast imaging at the “radiology” spatial scale, by simulating tissue structures seen in radiological images, namely, skin, adipose or dense tissue compartments, fibrous ligaments, and major ducts and blood vessels. We have extended this simulation to the microscopic (i.e., “pathology”) spatial scale, to allow for virtual rad-path correlation. Previously, we developed a manual simulation of adipose and dense tissue regions in pathology images, corresponding to a small region selected within a breast phantom simulated at the radiological scale. In this paper, we describe an automated simulation of adipocytes, epithelial and myoepithelial cells, collagen fibers, and fibroblasts. Adipocytes are simulated by recursive partitioning. Epithelial and myoepithelial cells are simulated radially around ductal or acinar lumen. Fibers and fibroblasts are simulated by an analogy with the electrostatic field. Our approach models the volumetric distributions of cells and various breast tissues, which allows the simulation of sequential pathology images at clinical inter-slice distances. The proposed simulation method has been evaluated by a clinical pathologist and medical physicists. The effect of the simulation approaches on the visual appearance of simulated pathology images has been evaluated.
Anthropomorphic software breast phantoms are generated by simulating breast anatomy. Virtual Clinical Trial (VCT) tools are developed for evaluating novel imaging modalities, based on anthropomorphic breast phantoms. Simulation of breast anatomical structures requires informed selection of parameters, which is crucial for the simulation realism. Our goal is to optimize the parameter selection based upon the analysis of clinical images.
Adipose compartments defined by Cooper’s ligaments significantly contribute to breast image texture (parenchymal pattern) which affects image interpretation and lesion detection. We have investigated the distribution and orientation of compartments segmented from CT images of a mastectomy specimen. Ellipsoidal fitting was applied to 205 segmented compartments, by matching the moments of inertia. The goodness-of-fit was measured by calculating Dice coefficients. Compartment size, shape, and orientation were characterized by estimating the volume, axis ratio, and Euler’s angles of fitted ellipsoids. Potential correlations between estimated parameters were tested.
We found that the adipose compartments are well approximated with ellipsoids (average Dice coefficient of 0.79). The compartment size is correlated with the barycenter-chest wall distance (r=0.235, p-value<0.001). The goodness-of-fit to ellipsoids is correlated to the compartment shape (r=0.344, p-value<0.001). The shape is also correlated with barycenter coordinates. The compartment orientation is correlated to their size (Euler angle α: r=0.188, p-value=0.007; angle β: r=0.156, p-value=0.025) and the barycenter-chest wall distance (r=0.159, p-value=0.023). These results from the characterization of adipose compartments and the observed correlations could help improve the realism of simulated breast anatomy.
Computer simulation of breast anatomy plays a crucial role in virtual clinical trials (VCTs) for preclinical optimization of breast imaging systems. Software breast phantoms provide ground truth about tissue distribution and flexibility to cover anatomical variations. We have experience with designing software phantoms based upon recursive partitioning using octrees; these phantoms simulate tissue compartments and fibrous ligaments, which contribute to the parenchymal texture. Realistic simulation critically affects the image quality and the VCT accuracy. Our simulation method may result in artifacts (bumps and dents) due to prematurely stopped partitioning of octrees. These artifacts compromise the image quality by reducing ligament smoothness and distorting parenchymal texture. In this study, we discuss the phenomenology of the artifacts and propose utilization of a spherical approximation of cubes corresponding to the octree nodes, to assess minimal and maximal distance from a cube to a median surface of the ligament. We demonstrate that the proposed technique is complementary to our earlier method proposed to improve smoothness of simulated Cooper’s ligaments surface. We show that the proposed technique leads to observable changes in simulated phantom projections. The effect of the computational overhead introduced by the proposed method on the simulation time may be compensated by an efficient implementation. The proposed method may be also applied to the simulation of quasi-planar structures in other organs and (biological or non-biological) domains.
Virtual clinical trials (VCTs) were introduced as a preclinical alternative to clinical imaging trials, and for the evaluation of breast imaging systems. Realism in computer models of breast anatomy (software phantoms), critical for VCT performance, can be improved by optimizing simulation parameters based on the analysis of clinical images. We optimized the simulation to improve the realism of simulated tissue compartments, defined by the breast Cooper’s ligaments. We utilized the anonymized, previously acquired CT images of a mastectomy specimen to manually segment 205 adipose compartments. We generated 1,440 anthropomorphic breast phantoms based on octree recursive partitioning. These phantoms included variations of simulation parameters—voxel size, number of compartments, percentage of dense tissue, and shape and orientation of the compartments. We compared distributions of the compartment volumes in segmented CT images and phantoms using Kolmogrov-Smirnov (KS) distance, Kullback-Leibler (KL) divergence and a novel distance metric (based on weighted sum of distribution descriptors differences). We identified phantoms with the size distributions closest to CT images. For example, KS resulted in the phantom with 1000 compartments, ligament thickness of 0.4 mm and skin thickness of 12 mm. We applied multilevel analysis of variance (ANOVAN) to these distance measures to identify parameters that most significantly influence the simulated compartment size distribution. We have demonstrated an efficient method for the optimization of phantom parameters to achieve realistic distribution of adipose compartment size. The proposed methodology could be extended to other phantom parameters (e.g., ligaments and skin thicknesses), to further improve realism of the simulation and VCTs.
Anthropomorphic breast phantoms are important tools for a wide range of tasks including pre-clinical validation of novel imaging techniques. In order to improve the realism in the phantoms, assessment of simulated anatomical structures is crucial. Thickness of simulated Cooper’s ligaments influences the percentage of dense tissue, as well as qualitative and quantitative properties of simulated images.
We introduce three methods (2-dimensional watershed, 3-dimensional watershed, and facet counting) to assess the thickness of the simulated Cooper’s ligaments in the breast phantoms. For the validation of simulated phantoms, the thickness of ligaments has been measured and compared with the input thickness values. These included a total of 64 phantoms with nominal ligament thicknesses of 200, 400, 600, and 800 μm.
The 2-dimensional and 3-dimensional watershed transformations were performed to obtain the median skeleton of the ligaments. In the 2-dimensional watershed, the median skeleton was found cross-section by cross-section, while the skeleton was found for the entire 3-dimensional space in the 3-dimensional watershed. The thickness was calculated by taking the ratio of the total volume of ligaments and the volume of median skeleton. In the facet counting method, the ligament thickness was estimated as a ratio between estimated ligaments’ volume and average ligaments’ surface area.
We demonstrated that the 2-dimensional watershed technique overestimates the ligament thickness. Good agreement was found between the facet counting technique and the 3-dimensional watershed for assessing thickness. The proposed techniques are applicable for ligaments’ thickness estimation on clinical breast images, provided segmentation of Cooper’s ligaments has been performed.
Computer simulation of breast anatomy is an essential component of Virtual Clinical Trials, a preclinical approach to validate breast imaging systems. Realism of breast phantoms affects simulation studies and their acceptance among researchers. Previously, we developed a simulation of tissue compartments defined by the hierarchy of Cooper’s ligaments, based upon recursive partitioning using octrees. In this work, we optimize the simulation parameters to represent realistically the breast subcutaneous and retromammary tissue regions. As seen in clinical images, the subcutaneous and retromammary regions contain predominantly adipose tissue organized into relatively large compartments, as opposed to the predominantly glandular breast interior. To mimic such organization, we divided the phantom volume into “subcutaneous”, “retromammary”, and “interior” regions. Within each region, parameters controlling the size and orientation of tissue compartments were selected separately. In this preliminary study, we varied parameter values and calculated the corresponding average compartment volume in each region. The proposed method was evaluated using anatomic descriptors at both radiological and pathological spatial scales. We simulated the subcutaneous region as spanning 20% of the breast diameter, comparable to published analysis of breast CT images. We simulated tissue compartments with the average volume of 0.94 cm3, 0.89 cm3 and 0.31 cm3 in the subcutaneous, retromammary and interior regions, respectively. Those average volumes match within 12% the values reported from histological analysis. Future evaluation will include a comparison of simulated and clinical parenchymal descriptors. The proposed method will be extended to automate the parameter optimization, and simulate detailed spatial variation, to further improve the realism.
Anthropomorphic software breast phantoms have been introduced as a tool for quantitative validation of breast imaging
systems. Efficacy of the validation results depends on the realism of phantom images. The recursive partitioning
algorithm based upon the octree simulation has been demonstrated as versatile and capable of efficiently generating large
number of phantoms to support virtual clinical trials of breast imaging.
Previously, we have observed specific artifacts, (here labeled “dents”) on the boundaries of simulated Cooper’s
ligaments. In this work, we have demonstrated that these “dents” result from the approximate determination of the
closest simulated ligament to an examined subvolume (i.e., octree node) of the phantom. We propose a modification of
the algorithm that determines the closest ligament by considering a pre-specified number of neighboring ligaments
selected based upon the functions that govern the shape of ligaments simulated in the subvolume.
We have qualitatively and quantitatively demonstrated that the modified algorithm can lead to elimination or reduction
of dent artifacts in software phantoms. In a proof-of concept example, we simulated a 450 ml phantom with 333
compartments at 100 micrometer resolution. After the proposed modification, we corrected 148,105 dents, with an
average size of 5.27 voxels (5.27nl). We have also qualitatively analyzed the corresponding improvement in the
appearance of simulated mammographic images. The proposed algorithm leads to reduction of linear and star-like
artifacts in simulated phantom projections, which can be attributed to dents. Analysis of a larger number of phantoms is
Anthropomorphic software breast phantoms have been utilized for preclinical quantitative validation of breast imaging
systems. Efficacy of the simulation-based validation depends on the realism of phantom images. Anatomical
measurements of the breast tissue, such as the size and distribution of adipose compartments or the thickness of Cooper’s
ligaments, are essential for the realistic simulation of breast anatomy. Such measurements are, however, not readily
available in the literature. In this study, we assessed the statistics of adipose compartments as visualized in CT images of
a total mastectomy specimen. The specimen was preserved in formalin, and imaged using a standard body CT protocol
and high X-ray dose. A human operator manually segmented adipose compartments in reconstructed CT images using
ITK-SNAP software, and calculated the volume of each compartment. In addition, the time needed for the manual
segmentation and the operator’s confidence were recorded. The average volume, standard deviation, and the probability
distribution of compartment volumes were estimated from 205 segmented adipose compartments. We also estimated the
potential correlation between the segmentation time, operator’s confidence, and compartment volume. The statistical
tests indicated that the estimated compartment volumes do not follow the normal distribution. The compartment volumes
are found to be correlated with the segmentation time; no significant correlation between the volume and the operator
confidence. The performed study is limited by the mastectomy specimen position. The analysis of compartment volumes
will better inform development of more realistic breast anatomy simulation.
An automated method has been developed to insert realistic clusters of simulated microcalcifications (MCs) into
computer models of breast anatomy. This algorithm has been developed as part of a virtual clinical trial (VCT) software
pipeline, which includes the simulation of breast anatomy, mechanical compression, image acquisition, image
processing, display and interpretation. An automated insertion method has value in VCTs involving large numbers of
images. The insertion method was designed to support various insertion placement strategies, governed by probability
distribution functions (pdf). The pdf can be predicated on histological or biological models of tumor growth, or
estimated from the locations of actual calcification clusters. To validate the automated insertion method, a 2-AFC
observer study was designed to compare two placement strategies, undirected and directed. The undirected strategy
could place a MC cluster anywhere within the phantom volume. The directed strategy placed MC clusters within
fibroglandular tissue on the assumption that calcifications originate from epithelial breast tissue. Three radiologists were
asked to select between two simulated phantom images, one from each placement strategy. Furthermore, questions were
posed to probe the rationale behind the observer’s selection. The radiologists found the resulting cluster placement to be
realistic in 92% of cases, validating the automated insertion method. There was a significant preference for the cluster to
be positioned on a background of adipose or mixed adipose/fibroglandular tissues. Based upon these results, this
automated lesion placement method will be included in our VCT simulation pipeline.
Images derived from a “phantom” are useful for characterizing the performance of imaging systems. In particular, the modulation transfer properties of imaging detectors are traditionally assessed by physical phantoms consisting of an edge. More recently researchers have come to realize that quantifying the effects of object variability can also be accomplished with phantoms in modalities such as breast imaging where anatomical structure may be the principal limitation in performance. This has driven development of virtual phantoms that can be used in simulation environments. In breast imaging, several such phantoms have been proposed. In this work, we analyze non-Gaussian statistical properties of virtual phantoms, and compare them to similar statistics from a database of breast images. The virtual phantoms assessed consist of three classes. The first is known as clustered-blob lumpy backgrounds. The second class is “binarized” textures which typically apply some sort of threshold to a stochastic 3D texture intended to represent the distribution of adipose and glandular tissue in the breast. The third approach comes from efforts at the University of Pennsylvania to directly simulate the 3D anatomy of the breast. We use Laplacian fractional entropy (LFE) as a measure of the non-Gaussian statistical properties of each simulation. Our results show that the simulation approaches differ considerably in LFE with very low scores for the clustered-blob lumpy background to very high values for the UPenn phantom. These results suggest that LFE may have value in developing and tuning virtual phantom simulation procedures.
Proc. SPIE. 8857, Signal and Data Processing of Small Targets 2013
KEYWORDS: Infrared imaging, Principal component analysis, Databases, Video, Surveillance, Signal processing, Reconstruction algorithms, Optimization (mathematics), Evolutionary algorithms, RGB color model
Detection of unusual trajectories of moving objects (e.g., people, automobiles, etc.) is an important problem
in many civilian and military surveillance applications. In this work, we propose a multi-objective evolutionary
algorithms and rough sets-based approach that breaks down 2-dimensional trajectories into a set
of additive components, which then can be used to build a classifier capable of recognizing typical, but yet
unseen trajectories, and identifying those that seem suspicious.
Software breast phantoms have been developed for use in evaluation of novel breast imaging systems. Software
phantoms are flexible allowing the simulation of wide variations in breast anatomy, and provide ground truth for the
simulated tissue structures. Different levels of phantom realism are required depending on the intended application.
Realistic simulation of dense (fibroglandular) tissue is of particular importance; the properties of dense tissue – breast
percent density and the spatial distribution – have been related to the risk of breast cancer. In this work, we have
compared two methods for simulation of dense tissue distribution in a software breast phantom previously developed at
the University of Pennsylvania. The methods compared are: (1) the previously used Gaussian distribution centered at
the phantom nipple point, and (2) the proposed combination of two Beta functions, one modeling the dense tissue
distribution along the chest wall-to-nipple direction, and the other modeling the radial distribution in each coronal
section of the phantom. Dense tissue distributions obtained using these methods have been compared with distributions
reported in the literature estimated from the analysis of breast CT images. Qualitatively, the two methods produced
rather similar dense tissue distributions. The simulation based upon the use of Beta functions provides more control
over the simulated distributions through the selection of the various Beta function parameters. Both methods showed
good agreement to the clinical data, suggesting both provide a high level of realism.
A novel breast image registration method is proposed to obtain a composite mammogram from several images with
partial breast coverage, for the purpose of accurate breast density estimation. The breast percent density estimated as a
fractional area occupied by fibroglandular tissue has been shown to be correlated with breast cancer risk. Some
mammograms, however, do not cover the whole breast area, which makes the interpretation of breast density estimates
ambiguous. One solution is to register and merge mammograms, yielding complete breast coverage. Due to elastic
properties of breast tissue and differences in breast positioning and deformation during the acquisition of individual
mammograms, the use of linear transformations does not seem appropriate for mammogram registration. Non-linear
transformations are limited by the changes in the mammographic projections pixel intensity with different positions of
the focal spot. We propose a novel method based upon non-linear local affine transformations. Initially, pairs of feature
points are manually selected and used to compute the best fit affine transformation in their small neighborhood. Finally, Shepherd interpolation is employed to compute affine transformations for the rest of the image area. The pixel values in the composite image are assigned using bilinear interpolation. Preliminary results with clinical images show a good match of breast boundaries, providing an increased coverage of breast tissue. The proposed transformation is continued and can be controlled locally. Moreover, the method is converging to the ground truth deformation if the paired feature points are evenly distributed and its number large enough.
Recent advances in high-resolution 3D breast imaging, namely, digital breast tomosynthesis and dedicated breast CT,
have enabled detailed analysis of the shape and distribution of anatomical structures in the breast. Such analysis is
critically important, since the projections of breast anatomical structures make up the parenchymal pattern in clinical
images which can mask the existing abnormalities or introduce false alarms; the parenchymal pattern is also correlated
with the risk of cancer. As a first step towards the shape analysis of anatomical structures in the breast, we have
analyzed an anthropomorphic software breast phantom. The phantom generation is based upon the recursive splitting of
the phantom volume using octrees, which produces irregularly shaped tissue compartments, qualitatively mimicking the
breast anatomy. The shape analysis was performed by fitting ellipsoids to the simulated tissue compartments. The
ellipsoidal semi-axes were calculated by matching the moments of inertia of each individual compartment and of an
ellipsoid. The distribution of Dice coefficients, measuring volumetric overlap between the compartment and the
corresponding ellipsoid, as well as the distribution of aspect ratios, measuring relative orientations of the ellipsoids, were
used to characterize various classes of phantoms with qualitatively distinctive appearance. A comparison between input
parameters for phantom generation and the properties of fitted ellipsoids indicated the high level of user control in the
design of software breast phantoms. The proposed shape analysis could be extended to clinical breast images, and used
to inform the selection of simulation parameters for improved realism.
A modification to our previous simulation of breast anatomy is proposed, in order to improve the quality of
simulated projections generated using software breast phantoms. Anthropomorphic software breast phantoms have
been used for quantitative validation of breast imaging systems. Previously, we developed a novel algorithm for
breast anatomy simulation, which did not account for the partial volume (PV) of various tissues in a voxel; instead,
each phantom voxel was assumed to contain single tissue type. As a result, phantom projection images displayed
notable artifacts near the borders between regions of different materials, particularly at the skin-air boundary. These
artifacts diminished the realism of phantom images. One solution is to simulate smaller voxels. Reducing voxel
size, however, extends the phantom generation time and increases memory requirements. We achieved an
improvement in image quality without reducing voxel size by the simulation of PV in voxels containing more than
one simulated tissue type. The linear x-ray attenuation coefficient of each voxel is calculated by combining
attenuation coefficients proportional to the voxel subvolumes occupied by the various tissues. A local planar
approximation of the boundary surface is employed, and the skin volume in each voxel is computed by
decomposition into simple geometric shapes. An efficient encoding scheme is proposed for the type and proportion
of simulated tissues in each voxel. We illustrate the proposed methodology on phantom slices and simulated
mammographic projections. Our results show that the PV simulation has improved image quality by reducing
A roadmap has been proposed to optimize the simulation of breast anatomy by parallel implementation, in order to
reduce the time needed to generate software breast phantoms. The rapid generation of high resolution phantoms is
needed to support virtual clinical trials of breast imaging systems. We have recently developed an octree-based
recursive partitioning algorithm for breast anatomy simulation. The algorithm has good asymptotic complexity;
however, its current MATLAB implementation cannot provide optimal execution times. The proposed roadmap for
efficient parallelization includes the following steps: (i) migrate the current code to a C/C++ platform and optimize it
for single-threaded implementation; (ii) modify the code to allow for multi-threaded CPU implementation; (iii) identify
and migrate the code to a platform designed for multithreaded GPU implementation. In this paper, we describe our
results in optimizing the C/C++ code for single-threaded and multi-threaded CPU implementations. As the first step of
the proposed roadmap we have identified a bottleneck component in the MATLAB implementation using MATLAB's
profiling tool, and created a single threaded CPU implementation of the algorithm using C/C++'s overloaded operators
and standard template library. The C/C++ implementation has been compared to the MATLAB version in terms of
accuracy and simulation time. A 520-fold reduction of the execution time was observed in a test of phantoms with 50-
400 μm voxels. In addition, we have identified several places in the code which will be modified to allow for the next
roadmap milestone of the multithreaded CPU implementation.
Recent research in motion detection has shown that various outlier detection methods could be used for efficient
detection of small moving targets. These algorithms detect moving objects as outliers in a properly defined attribute
space, where outlier is defined as an object distinct from the objects in its neighborhood. In this paper, we compare the
performance of two incremental outlier detection algorithms, namely the incremental connectivity-based outlier factor
and the incremental local outlier factor to modified Stauffer-Grimson algorithm. Each video sequence is represented
with spatial-temporal blocks extracted from the raw video. Principal component analysis (PCA) is applied on these
blocks in order to reduce the dimensionality of extracted data. Extensive experiments performed on several data sets,
including infrared sequences from OSU Thermal Pedestrian Database repository, and data collected at Delaware State
University from FLIR Systems PTZ cameras have shown promising results in using outlier detection for detection of
small moving targets.
Proc. SPIE. 7445, Signal and Data Processing of Small Targets 2009
KEYWORDS: Data mining, Detection and tracking algorithms, Data modeling, Databases, Video, Image analysis, Video surveillance, Improvised explosive devices, Expectation maximization algorithms, Scalable video coding
Detection of unusual trajectories of moving objects can help in identifying suspicious activity on convoy routes and thus
reduce casualties caused by improvised explosive devices. In this paper, using video imagery we compare efficiency of
various techniques for incremental outlier detection on detecting unusual trajectories on simulated and real-life data
obtained from SENSIAC database. Incremental outlier detection algorithms that we consider in this paper include
incremental Support Vector Classifier (incSVC), incremental Local Outlier Factor (incLOF) algorithm and incremental
Connectivity Outlier Factor (incCOF) algorithm. Our experiments performed on ground truth trajectory data indicate that
incremental LOF algorithm can provide better detection of unusual trajectories in comparison to other examined
Fast Fourier transform spectroscopy has proved to be a powerful method for study of the secondary structure of proteins
since peak positions and their relative amplitude are affected by the number of hydrogen bridges that sustain this
secondary structure. However, to our best knowledge, the method has not been used yet for identification of proteins
within a complex matrix like a blood sample. The principal reason is the apparent similarity of protein infrared spectra
with actual differences usually masked by the solvent contribution and other interactions. In this paper, we propose a
novel machine learning based method that uses protein spectra for classification and identification of such proteins
within a given sample. The proposed method uses principal component analysis (PCA) to identify most important linear
combinations of original spectral components and then employs support vector machine (SVM) classification model
applied on such identified combinations to categorize proteins into one of given groups. Our experiments have been
performed on the set of four different proteins, namely: Bovine Serum Albumin, Leptin, Insulin-like Growth Factor 2
and Osteopontin. Our proposed method of applying principal component analysis along with support vector machines
exhibits excellent classification accuracy when identifying proteins using their infrared spectra.
Although a tremendous effort has been made to perform a reliable analysis of images and videos in the past fifty years, the reality is that one cannot rely 100% on the analysis results. The only exception is applications in controlled environments as dealt in machine vision, where closed world assumptions apply. However, in general, one has to deal with an open world, which means that content of images may significantly change, and it seems impossible to predict all possible changes. For example, in the context of surveillance videos, the light conditions may suddenly fluctuate in parts of images only, video compression or transmission artifacts may occur, a wind may cause a stationary camera to tremble, and so on. The problem is that video analysis has to be performed in order to detect content changes, but such analysis may be unreliable due to the changes, and thus fail to detect the changes and lead to "vicious cycle".
The solution pursuit in this paper is to monitor the reliability of the computed features by analyzing their general properties. We consider statistical properties of feature value distributions as well as temporal properties. Our main strategy is to estimate the feature properties when the features are reliable computed, so that any set of features that does not have these properties is detected as being unreliable. This way we do not perform any direct content analysis, but instead perform analysis of feature properties related to their reliability.
We propose partitioning-based methods to facilitate the classification of 3-D binary image data sets of regions of interest (ROIs) with highly non-uniform distributions. The first method is based on recursive dynamic partitioning of a 3-D volume into a number of 3-D hyper-rectangles. For each hyper-rectangle, we consider, as a potential attribute, the number of voxels (volume elements) that belong to ROIs. A hyper-rectangle is partitioned only if the corresponding attribute does not have high discriminative power, determined by statistical tests, but it is still sufficiently large for further splitting. The final discriminative hyper-rectangles form new attributes that are further employed in neural network classification models. The second method is based on maximum likelihood employing non-spatial (k-means) and spatial DBSCAN clustering algorithms to estimate the parameters of the underlying distributions. The proposed methods were experimentally evaluated on mixtures of Gaussian distributions, on realistic lesion-deficit data generated by a simulator conforming to a clinical study, and on synthetic fractal data. Both proposed methods have provided good classification on Gaussian mixtures and on realistic data. However, the experimental results on fractal data indicated that the clustering-based methods were only slightly better than random guess, while the recursive partitioning provided significantly better classification accuracy.