Lawrence A. Ray

Performance of topological texture features to classify fibrotic interstitial lung disease patterns.

PURPOSEnTopological texture features were compared in their ability to classify honeycombing, a morphological pattern that is considered indicative for the presence of fibrotic interstitial lung disease in high-resolution computed tomography (HRCT) images.nnnMETHODSnFor 14 patients with known occurrence of honeycombing, a stack of 70 axial, lung kernel reconstructed images was acquired from HRCT chest exams. A set of 964 regions of interest of both healthy and pathological (356) lung tissue was identified by an experienced radiologist. Texture features were extracted using statistical features (Stat), six properties calculated from gray-level co-occurrence matrices (GLCMs), Minkowski dimensions (MDs), and three Minkowski functionals (MFs) (e.g., MF.Euler). A naïve Bayes (NB) and k-nearest-neighbor (k-NN) classifier, a multilayer radial basis functions network (RBFN), and a support vector machine with a radial basis function (SVMrbf) kernel were optimized in a tenfold cross-validation for each texture vector, and the classification accuracy was calculated on independent test sets as a quantitative measure of automated tissue characterization. A Wilcoxon signed-rank test was used to compare two accuracy distributions and the significance thresholds were adjusted for multiple comparisons by the Bonferroni correction.nnnRESULTSnThe best classification results were obtained by the MF features, which performed significantly better than all the standard Stat, GLCM, and MD features (p < 0.001) for both classifiers. The highest accuracies were found for MF.Euler (93.6%, 94.9%, 94.2%, and 95.0% for NB, k-NN, RBFN, and SVMrbf, respectively). The best groups of standard texture features were a Stat and GLCM (homogeneity) feature set (up to 91.8%).nnnCONCLUSIONSnThe results indicate that advanced topological texture features derived from MFs can provide superior classification performance in computer-assisted diagnosis of fibrotic interstitial lung disease patterns when compared to standard texture analysis methods.

Classification of Interstitial Lung Disease Patterns with Topological Texture Features

Topological texture features were compared in their ability to classify morphological patterns known as honeycombing that are considered indicative for the presence of fibrotic interstitial lung diseases in high-resolution computed tomography (HRCT) images. For 14 patients with known occurrence of honey-combing, a stack of 70 axial, lung kernel reconstructed images were acquired from HRCT chest exams. A set of 241 regions of interest of both healthy and pathological (89) lung tissue were identified by an experienced radiologist. Texture features were extracted using six properties calculated from gray-level co-occurrence matrices (GLCM), Minkowski Dimensions (MDs), and three Minkowski Functionals (MFs, e.g. MF.euler). A k-nearest-neighbor (k-NN) classifier and a Multilayer Radial Basis Functions Network (RBFN) were optimized in a 10-fold cross-validation for each texture vector, and the classification accuracy was calculated on independent test sets as a quantitative measure of automated tissue characterization. A Wilcoxon signed-rank test was used to compare two accuracy distributions and the significance thresholds were adjusted for multiple comparisons by the Bonferroni correction. The best classification results were obtained by the MF features, which performed significantly better than all the standard GLCM and MD features (p < 0.005) for both classifiers. The highest accuracy was found for MF.euler (97.5%, 96.6%; for the k-NN and RBFN classifier, respectively). The best standard texture features were the GLCM features homogeneity (91.8%, 87.2%) and absolute value (90.2%, 88.5%). The results indicate that advanced topological texture features can provide superior classification performance in computer-assisted diagnosis of interstitial lung diseases when compared to standard texture analysis methods.

Texture feature ranking with relevance learning to classify interstitial lung disease patterns

OBJECTIVEnThe generalized matrix learning vector quantization (GMLVQ) is used to estimate the relevance of texture features in their ability to classify interstitial lung disease patterns in high-resolution computed tomography images.nnnMETHODOLOGYnAfter a stochastic gradient descent, the GMLVQ algorithm provides a discriminative distance measure of relevance factors, which can account for pairwise correlations between different texture features and their importance for the classification of healthy and diseased patterns. 65 texture features were extracted from gray-level co-occurrence matrices (GLCMs). These features were ranked and selected according to their relevance obtained by GMLVQ and, for comparison, to a mutual information (MI) criteria. The classification performance for different feature subsets was calculated for a k-nearest-neighbor (kNN) and a random forests classifier (RanForest), and support vector machines with a linear and a radial basis function kernel (SVMlin and SVMrbf).nnnRESULTSnFor all classifiers, feature sets selected by the relevance ranking assessed by GMLVQ had a significantly better classification performance (p<0.05) for many texture feature sets compared to the MI approach. For kNN, RanForest, and SVMrbf, some of these feature subsets had a significantly better classification performance when compared to the set consisting of all features (p<0.05).nnnCONCLUSIONnWhile this approach estimates the relevance of single features, future considerations of GMLVQ should include the pairwise correlation for the feature ranking, e.g. to reduce the redundancy of two equally relevant features.

FAST FLUID REGISTRATION WITH DIRICHLET BOUNDARY CONDITIONS: A TRANSFORM-BASED APPROACH

Fluid registration is an example of a nonrigid image registration algorithm that uses a deformation field to define the transformation between two images. The velocity of the deformation field is governed by the Navier-Lame equations, which can be discretized and solved numerically via fixed-point iteration. The fixed-point iteration generates a succession of linear PDE systems, which can be solved quickly via discrete Fourier transform (DFT) techniques, as shown in the prior art. The major drawback of this approach is that it is only applicable when the boundary conditions of the velocity field are assumed to be periodic. This paper shows that by considering the adjoint of the Navier-Lame operator, the succession of linear PDE systems can be solved quickly via discrete sine transform (DST) techniques, generating velocity fields that satisfy Dirichlet boundary conditions (where the velocities are zero on the boundaries)

3-D examination of dental fractures with minimum user intervention

We developed a novel, powerful segmentation algorithm and an intuitive 3-D visualization tool for the examination of root fractures with minimum user intervention. The application computes and displays a suitable oblique orientation on a selected tooth by placing at least two splines (inside and outside of the tooth) in just one slice of the volume. Next, it allows the user to scroll through the volume, slice-by-slice in parallel to the plane, or to examine the tooth by changing the orientation of a 3-D object plane (called a virtual bitewing), which is placed, at the same time, in a volume rendition. Both the root canal and the root fracture are highlighted during the examination phase. Doctors (end users) are in control to quickly and confidently examine root fractures in 3-D, for any given oblique orientation, without worrying about missing a selected tooth. We have designed and implemented these algorithms using the image foresting transform (IFT) technique for interactive tooth segmentation and used a multi-scale parameter search for automatic oblique orientation estimation.

Texture feature selection with relevance learning to classify interstitial lung disease patterns

The Generalized Matrix Learning Vector Quantization (GMLVQ) is used to estimate the relevance of texture features in their ability to classify interstitial lung disease patterns in high-resolution computed tomography (HRCT) images. After a stochastic gradient descent, the GMLVQ algorithm provides a discriminative distance measure of relevance factors, which can account for pairwise correlations between different texture features and their importance for the classification of healthy and diseased patterns. Texture features were extracted from gray-level co-occurrence matrices (GLCMs), and were ranked and selected according to their relevance obtained by GMLVQ and, for comparison, to a mutual information (MI) criteria. A k-nearest-neighbor (kNN) classifier and a Support Vector Machine with a radial basis function kernel (SVMrbf) were optimized in a 10-fold crossvalidation for different texture feature sets. In our experiment with real-world data, the feature sets selected by the GMLVQ approach had a significantly better classification performance compared with feature sets selected by a MI ranking.

2-D and 3-D Image Registration for Medical, Remote Sensing, and Industrial Applications .

This PDF file contains the editorial “Book Review: 2-D and 3-D Image Registration for Medical, Remote Sensing, and Industrial Applications” for JEI Vol. 14 Issue 03

3D printing for orthopedic applications: from high resolution cone beam CT images to life size physical models

With increasing resolution in image acquisition, the project explores capabilities of printing toward faithfully reflecting detail and features depicted in medical images. To improve safety and efficiency of orthopedic surgery and spatial conceptualization in training and education, this project focused on generating virtual models of orthopedic anatomy from clinical quality computed tomography (CT) image datasets and manufacturing life-size physical models of the anatomy using 3D printing tools. Beginning with raw micro CT data, several image segmentation techniques including thresholding, edge recognition, and region-growing algorithms available in packages such as ITK-SNAP, MITK, or Mimics, were utilized to separate bone from surrounding soft tissue. After converting the resulting data to a standard 3D printing format, stereolithography (STL), the STL file was edited using Meshlab, Netfabb, and Meshmixer. The editing process was necessary to ensure a fully connected surface (no loose elements), positive volume with manifold geometry (geometry possible in the 3D physical world), and a single, closed shell. The resulting surface was then imported into a “slicing” software to scale and orient for printing on a Flashforge Creator Pro. In printing, relationships between orientation, print bed volume, model quality, material use and cost, and print time were considered. We generated anatomical models of the hand, elbow, knee, ankle, and foot from both low-dose high-resolution cone-beam CT images acquired using the soon to be released scanner developed by Carestream, as well as scaled models of the skeletal anatomy of the arm and leg, together with life-size models of the hand and foot.

Detection of tooth fractures in CBCT images using attention index estimation

The attention index (𝜑) is a number from zero to one that indicates a possible fracture is detected inside a selected tooth. The higher the 𝜑 number, the greater the likelihood for needed attention in the visual examination. The method developed for the 𝜑 estimation extracts a connected component with image properties that are similar to those of a typical tooth fracture. That is, in cone-beam computed tomography (CBCT) images, a fracture appears as a dark canyon inside the tooth. In order to start the visual examination process, the method provides a plane across the geometric center of the suspicious fracture component, which maximizes the number of pixels from that component inside the plane. During visual examination, the user (doctor) can change plane orientations and locations, by manipulating the mouse toward different graphical elements that represent the plane on a 3-D rendition of the tooth, while the corresponding image of the plane is shown at its side. The visual examination aims at confirming or disproving the fracture-detection event. We have designed and implemented these algorithms using the image-foresting transform methodology.

Lesion classification on breast MRI through topological characterization of morphology over time

Morphological characterization of lesions on dynamic breast MRI exams through texture analysis has typically involved the computation of gray-level co-occurrence matrices (GLCM), which serve as the basis for second order statistical texture features. This study aims to characterize lesion morphology through the underlying topology and geometry with Minkowski Functionals (MF) and investigate the impact of using such texture features extracted dynamically over a time series in classifying benign and malignant lesions. 60 lesions (28 malignant & 32 benign) were identified and annotated by experienced radiologists on 54 breast MRI exams of female patients where histopathological reports were available prior to this investigation. 13 GLCM-derived texture features and 3 MF features were then extracted from lesion ROIs on all five post-contrast images. These texture features were combined into high dimensional texture feature vectors and used in a lesion classification task. A fuzzy k-nearest neighbor classifier was optimized using random sub-sampling cross-validation for each texture feature and the classification performance was calculated on an independent test set using the area under the ROC curve (AUC); AUC distributions of different features were compared using a Mann- Whitney U-test. The MF feature Area exhibited significantly improvements in classification performance (p<0.05) when compared to all GLCM-derived features while the MF feature Perimeter significantly outperformed 12 out of 13 GLCM features (p<0.05) in the lesion classification task. These results show that dynamic texture tracking of morphological characterization that relies on topological texture features can contribute to better lesion character classification.