Search

Simple Summary: In recent years, the prostate cancer histopathological description proposed by Gleason has emerged as a universal standard used for disease diagnosis and progression. Recently, a grading scheme on a point scale is based on Gleason patterns. Current scores are highly dependent on the expert urinary histopathologist and show a high level of variability among experts. To aid the clinician, we have developed deep learning models that provide a decision aid in identifying the primary cancer grade (dominant Gleason pattern). Histopathological classification in prostate cancer remains a challenge with high dependence on the expert practitioner. We develop a deep learning (DL) model to identify the most prominent Gleason pattern in a highly curated data cohort and validate it on an independent dataset. The histology images are partitioned in tiles (14,509) and are curated by an expert to identify individual glandular structures with assigned primary Gleason pattern grades. We use transfer learning and fine-tuning approaches to compare several deep neural network architectures that are trained on a corpus of camera images (ImageNet) and tuned with histology examples to be context appropriate for histopathological discrimination with small samples. In our study, the best DL network is able to discriminate cancer grade (GS3/4) from benign with an accuracy of 91%, F1-score of 0.91 and AUC 0.96 in a baseline test (52 patients), while the cancer grade discrimination of the GS3 from GS4 had an accuracy of 68% and AUC of 0.71 (40 patients). [ABSTRACT FROM AUTHOR]

• We proposed an Automated Segmentation Algorithm (ASA) to segment neurons and microglia cells in NeuN- and Iba1-stained images. • The proposed algorithm is used to automate the stereology task of counting cells. • ASA segments cells in Extended Depth of Field (EDF) images using a combination of several image processing techniques such as Gaussian Mixture Model (GMM), morphological operations, watershed segmentation, Voronoi diagrams and boundary smoothing in a novel approach. • The experiments for neuron counting, show a high correlation between the automated neuron counts and manual counts. • It has been observed that the automated microglia counts are more accurate than the manual counts on a (non-expert) data collectors when compared to the counts of an expert. • Moreover, the segmentation output of ASA is shown to be useful in training a deep learning neural network model (in ongoing work). • The results show that ASA can be reliably used to automate the time-consuming task of cell counting, and it can potentially be revised and extended to count other types of cells in mouse neocortex images. Collection of unbiased stereology data currently relies on relatively simple, low throughput technology developed in the mid-1990s. In an effort to improve the accuracy and efficiency of these integrated hardware-software-digital microscopy systems, we have developed an automatic segmentation algorithm (ASA) for automatic stereology counts using the unbiased optical fractionator method. Here we report on a series of validation experiments in which immunostained neurons (NeuN) and microglia (Iba1) were automatically counted in tissue sections through a mouse neocortex. In the first step, a minimum of 100 systematic-random z -axis image stacks (disector stacks) containing NeuN- and Iba1-immunostained cells were automatically collected using a software-controlled 3 axes (XYZ) stage motor. In the second step, each disector stack was converted to an extended depth of field (EDF) image in which each cell is shown at its optimal plane of focus. Third, individual neurons and microglia were segmented and the regional minimas were extracted and used as seed regions for cells in a watershed segmentation algorithm. Finally, the unbiased disector frame and counting rules were used to make unbiased parameter estimates for neurons and microglia cells. The results for both NeuN neurons and Iba1 microglia were compared to manual counts made by a moderately experienced data collector from the same disector stacks. The final results show lower error rates for counts of Iba1-immunostained microglia cells than for quantifying NeuN-immunostained neurons, most likely due to less three-dimensional overlapping of Iba1 cells. We report that the throughput efficiency of using ASA to automatically annotate images of Iba1 microglia is more than five times greater than that of manual stereology counts of the same sections. Moreover, we show that ASA is significantly more accurate in counting microglia cells than a moderately experienced data collector (about 10% higher overall accuracy) when both were compared to counts by an expert neurohistologist. Thus, the ASA method applied to EDF images from disector stacks can be extremely useful to automate and increase the accuracy of cell counts, which could be especially helpful and cost-effective when expert help is not available. Another potential use of our ASA approach is to generate unsupervised ground truth as an efficient alternative to manual annotation for training deep learning models, as shown in our ongoing work. [ABSTRACT FROM AUTHOR]

Highlights • We proposed a framework to automate the manual process of stereological point counting using a novel and robust segmentation method that segmented immunostained neuron cells in the rat substantia nigra images. • Compared the manual counts to automatic counts and found that there is a high correlation between them. • In all but one case the difference between the manual and automated count was about or less than 5% of the manual count. • Most of the important parameters in the framework are set adaptively for each image separately, and therefore, the whole framework is resistant to minor to moderate changes in image intensity and staining process. • The findings suggest that the framework can be used reliably to automate the time consuming manual process of point counting in rat brain images. Abstract The use of unbiased stereology to quantify structural parameters such as mean cell and nuclear size (area and volume) can be useful for a wide variety of biological studies. Here we propose a novel segmentation framework using an Active Contour Model to automate the collection of stereology from stained cells and other objects in tissue sections. This approach is demonstrated for stained brain sections from young adult Fischer 344 rats. Animals were perfused in-vivo with 4% paraformaldehyde and sectioned by frozen microtomy at an instrument setting of 40 μm. For each rat brain, a systematic-random set of sections through the entire substantia nigra pars compacta (SN) were immunostained to reveal tyrosine hydroxylase (TH)-immunopositive neurons. The novel framework applied an active contour (modified balloon snake) model with non-constant balloon force to automatically segment and quantify neuronal cell bodies by stereological point counting (SPC). Several contours were initialized in the image and based on the contour fit after 200 iterations classified as immunopositive (signal) or background contours in a sequential manner. Cell contours were determined in four steps based on several criteria, e.g., area of contour, dispersion measure, and degree of overlap. The image was automatically segmented according to the final contours. Using a point grid automatically generated at systematic-random orientations over the images, points hitting the segmented neural cell bodies were automatically counted. The final values from the automatic framework were compared with findings for ground truth (manual SPC). The results of this study show a strong agreement between data collected by the automatic framework and the ground truth (R 2 ≥ 0.95) with a 5× gain in time efficiency for the automatic SPC. These findings give strong support for future applications of pattern recognition for assessing stereological parameters of biological objects identified by high signal:noise stains. [ABSTRACT FROM AUTHOR]

Highlights • Conventional unbiased stereology suffers from low accuracy and poor inter-rater reliability. • An unsupervised algorithm (ASA) can mitigate human effort on data labeling. • CNN learns discriminant and powerful features to segment cells. • Cell count using deep learning based unbiased stereology showed high accuracy and reproducibility. Abstract In recent decades stereology-based studies have played a significant role in understanding brain aging and developing novel drug discovery strategies for treatment of neurological disease and mental illness. A major obstacle to further progress in a wide range of neuroscience sub-disciplines remains the lack of high-throughput technology for stereology analyses. Though founded on methodologically unbiased principles, commercially available stereology systems still rely on well-trained humans to manually count hundreds of cells within each region of interest (ROI). Even for a simple study with 10 controls and 10 treated animals, cell counts typically require over a month of tedious labor and high costs. Furthermore, these studies are prone to errors and poor reproducibility due to human factors such as subjectivity, variable training, recognition bias, and fatigue. Here we propose a deep neural network-stereology combination to automatically segment and estimate the total number of immunostained neurons on tissue sections. Our three-step approach consists of (1) creating extended-depth-of-field (EDF) images from z-stacks of images (disector stacks); (2) applying an adaptive segmentation algorithm (ASA) to label stained cells in the EDF images (i.e., create masks) for training a convolutional neural network (CNN); and (3) use the trained CNN model to automatically segment and count the total number of cells in test disector stacks using the optical fractionator method. The automated stereology approach shows less than 2% error and over 5× greater efficiency compared to counts by a trained human, without the subjectivity, tedium, and poor precision associated with conventional stereology. [ABSTRACT FROM AUTHOR]

We propose a framework to detect and segment nuclei and segment overlapping cytoplasm in cervical cytology images. This is a challenging task due to folded cervical cells with spurious edges, poor contrast of cytoplasm and presence of neutrophils and artifacts. The algorithm segments nuclei and cell clumps in extended depth of field (EDF) images and uses volume images to segment overlapping cytoplasm. The boundaries are first approximated by a defined similarity metric and are refined in two steps by reducing concavity, iterative smoothing and outliers removal. We evaluated our framework on two public datasets provided in the first and second overlapping cervical cell segmentation challenges (ISBI 2014 and 2015). The results show that our method outperforms other state-of-the-art algorithms on both datasets. The results on the ISBI 2014 dataset show that our method missed less than 5% of cells when the pairwise cell overlapping degree was not higher than 0.3 and it missed only 7% of cells on average in a dataset of 810 synthetic images with 4860 (overlapping) cells. On the same dataset, it outperforms other state-of-the-art methods in nucleus detection with precision 0.961 and recall 0.933. The results on the ISBI 2015 dataset containing real cervical EDF images show that our method misses around 20% of cells in EDF images where a segmentation is considered a miss if it has dice similarity coefficient not greater than 0.7. The 20% miss rate is around half of the miss rate of two other recent methods. [ABSTRACT FROM AUTHOR]

Stereology-based methods provide the current state-of-the-art approaches for accurate quantification of numbers and other morphometric parameters of biological objects in stained tissue sections. The advent of artificial intelligence (AI)-based deep learning (DL) offers the possibility of improving throughput by automating the collection of stereology data. We have recently shown that DL can effectively achieve comparable accuracy to manual stereology but with higher repeatability, improved throughput, and less variation due to human factors by quantifying the total number of immunostained cells at their maximal profile of focus in extended depth of field (EDF) images. In the first of two novel contributions in this work, we propose a semi-automatic approach using a handcrafted Adaptive Segmentation Algorithm (ASA) to automatically generate ground truth on EDF images for training our deep learning (DL) models to automatically count cells using unbiased stereology methods. This update increases the amount of training data, thereby improving the accuracy and efficiency of automatic cell counting methods, without a requirement for extra expert time. The second contribution of this work is a Multi-channel Input and Multi-channel Output (MIMO) method using a U-Net deep learning architecture for automatic cell counting in a stack of z -axis images (also known as disector stacks). This DL-based digital automation of the ordinary optical fractionator ensures accurate counts through spatial separation of stained cells in the z -plane, thereby avoiding false negatives from overlapping cells in EDF images without the shortcomings of 3D and recurrent DL models. The contribution overcomes the issue of under-counting errors with EDF images due to overlapping cells in the z -plane (masking). We demonstrate the practical applications of these advances with automatic disector-based estimates of the total number of NeuN-immunostained neurons in a mouse neocortex. In summary, this work provides the first demonstration of automatic estimation of a total cell number in tissue sections using a combination of deep learning and the disector-based optical fractionator method. [Display omitted] • Stereology methods are accurate for quantifying stained cells in tissue sections. • Previous deep learning (DL) methods for stereology cell counts rely on EDF images. • Possibility of under-counting error due to overlapping/ occluded cells in z -plane. • Disector-based multi-channel input multi-channel output (MIMO) DL method presented. • MIMO avoids under-counting by counting each cell only once at its best focal plane. [ABSTRACT FROM AUTHOR]

Abstract: Visual surveillance in the maritime domain has been explored for more than a decade. Although it has produced a number of working systems and resulted in a mature technology, surveillance has been restricted to the port facilities or areas close to the coast line assuming a fixed-camera scenario. This paper presents a novel algorithm for open-sea visual maritime surveillance. We explore a challenging situation with a forward-looking camera mounted on a buoy or other floating platform. The proposed algorithm detects, localizes, and tracks ships in the field of view of the camera. Specifically, developed algorithm is uniquely designed to handle rapidly moving camera. Its performance is robust in the presence of a random relatively-large camera motion. In the context of ship detection we developed a new horizon detection scheme for a complex maritime domain. The performance of our algorithm and its comprising elements is evaluated. Ship detection precision of 88% is achieved on a large dataset collected from a prototype system. [Copyright &y& Elsevier]

This paper develops a novel approach for fluid flow tracking and analysis. Specifically, the proposed algorithm is able to detect the traveling waves, compute the wave parameters and determine controlling film flow parameters for a liquid film flowing over a rotating disk. The input to the algorithm is an easily acquired video data. It is shown that under single light illumination, it is possible to track the specular portion of the reflected light on the traveling wave. Hence, wavy films can be tracked, and fluid flow parameters can be computed. The fluid flow parameters include wave velocities, wave inclination angles, and distances between consecutive waves. Once the parameters are computed, their accuracy is analyzed and compared with the solutions of the mathematical models based on the Navier-Stokes equations. The mathematical model predicts wave characteristics based on directly measured controlling parameters, such as disk rotation speed and fluid flow rate. It is shown that the calculated parameter values coincide with the predicted ones. The average computed parameters are within 5-10% of the predicted values. Next, the developed approach is generalized to model-based recovery of fluid flow controlling parameters: the rotation speed and the fluid-flow rate. The search in space for model parameters is performed to minimize the error between the flow characteristics predicted by the fluid dynamics model (e.g. distance between waves, wave inclination angles) and parameters recovered from video data. Results demonstrate that the speed of a disk and the flow rate, when compared to the ground truth available from direct observation, are recovered with the error less than 10%. [ABSTRACT FROM AUTHOR]

Common benchmark data sets, standardized performance metrics, and baseline algorithms have demonstrated considerable impact on research and development in a variety of application domains. These resources provide both consumers and developers of technology with a common framework to objectively compare the performance of different algorithms and algorithmic improvements. In this paper, we present such a framework for evaluating object detection and tracking in video: specifically for face, text, and vehicle objects. This framework includes the source video data, ground-truth annotations (along with guidelines for annotation), performance metrics, evaluation protocols, and tools including scoring software and baseline algorithms. For each detection and tracking task and supported domain, we developed a 50-clip training set and a 50-clip test set. Each data clip is approximately 2.5 minutes long and has been completely spatially/temporally annotated at the I-frame level. Each task/domain, therefore, has an associated annotated corpus of approximately 450,000 frames. The scope of such annotation is unprecedented and was designed to begin to support the necessary quantities of data for robust machine learning approaches, as well as a statistically significant comparison of the performance of algorithms. The goal of this work was to systematically address the challenges of object detection and tracking through a common evaluation framework that permits a meaningful objective comparison of techniques, provides the research community with sufficient data for the exploration of automatic modeling techniques, encourages the incorporation of objective evaluation into the development process, and contributes useful lasting resources of a scale and magnitude that will prove to be extremely useful to the computer vision research community for years to come. [ABSTRACT FROM AUTHOR]

Abstract: Parameters used in physical models for nonrigid and articulated motion analysis are often not known with high precision. It has been recognized that commonly used assumptions about the parameters may have adverse effect on modeling quality. In this paper, we present an efficient sensitivity analysis method to assess the impact of those assumptions by examining the model''s spatial response to parameter perturbation. Numerical experiments with a synthetic model and skin tissues show that: (1) normalized sensitivity distribution can help determine the relative importance of different parameters; (2) dimensional sensitivity is useful in the assessment of a particular parameter assumption; and (3) models are more sensitive at the locations of property discontinuity (heterogeneity). The formulation of the proposed sensitivity analysis method is general and can be applied to assessment of other types of assumptions, such as those related to nonlinearity and anisotropy. [Copyright &y& Elsevier]

A modeling approach is presented for quantitative burn scar assessment. Emphases are given to: 1) constructing a finite-element mode) from natural image features with an adaptive mesh and 2) quantifying the Young's modulus of scars using the finite-element model and regularization method, A set of natural point features is extracted from the images of burn patients. A Delaunay triangle mesh is then generated that adapts to the point features. A three-dimensional finite-element model is built on top of the mesh with the aid of range images providing the depth information. The Young's modulus of scars is quantified with a simplified regularization functional, assuming that the knowledge of the scar's geometry is available. The consistency between the relative elasticity index and the physician's rating based on the Vancouver scale (a relative scale used to rate burn scars) indicates that the proposed modeling approach has high potential for image-based quantitative burn scar assessment. [ABSTRACT FROM AUTHOR]

Presents a method to compute the egomotion of a range camera using the space envelope process. Description of a space envelope; Cases where the method fails; Steps in computing the egomotion.

We present a novel method for estimating motion parameters and point correspondences between 3D surfaces under small nonrigid motion. A vector point function is utilized as the motion parameter, called the displacement function. Differential-geometric changes of surfaces are then used in tracking small deformations. Discriminant (of first fundamental form), unit-normal and Gaussian curvature are the invariant differential-geometric parameters that have been utilized for nonrigid motion analysis.Tests were performed by generating nonrigid motion on a simulated data set to illustrate performance and accuracy of our algorithms. Experiments were then performed on a Cyberware range data sequence of facial motion. A total of 16 sets of facial motion images were used in our experiments, belonging to eight different persons, each having two facial expressions. We have demonstrated the correct point correspondence recovery by tracking features of the face during each facial expression and comparing against the manual tracking of feature points by different users. In addition, nonrigid motion segmentation and interpolation of intermediate frames of data were successfully performed on these images. We have also performed experiments on cardiac data in order to estimate the motion parameters related to the abnormality in cardiac motion. Two sets of volumetric CT data of the left ventricle of a dog''s heart in cardiac cycle were used in our experiments. All our experiments indicate that the system performs very well and proves to be extremely useful in other nonrigid motion analysis applications. [Copyright &y& Elsevier]

Presents a novel multiscale approach to recovery of nonrigid motion from sequences of registered intensity and range images. Description of the method; Application of the proposed method to motion analysis; Performance comparison for a number of different experiments involving elastic stretching.

We introduce a software framework called the Java Vision Toolkit (JVT) for teaching image processing and computer vision. The toolkit provides over 50 image operations and presents them to the user in a GUI that can render grayscale, color and 3D range images. The software is written in Java, enabling it to be integrated into HTML documents and interactive course materials. The framework is designed for extensibility using a source code template that supports the implementation of any new operation with a minimal amount of supporting code. For students, this framework encapsulates the GUI, file I/O and other trivial programming details and allows them the maximum amount of time to spend on understanding computer vision. We compare the JVT with other computer vision software frameworks that are used for teaching and research. We also discuss the use of the JVT in an undergraduate image processing course at the University of South Florida. [ABSTRACT FROM AUTHOR]

Deals with a study which evaluated edge detector performance through the use of the structure from motion task. Information on edge detectors; Functionality of a structure from motion algorithm; Results and conclusion.

Presents a study which proposed algorithms for nonrigid motion tracking. Importance of motion analysis in computer vision; Conclusions.

In this paper we propose a method for increasing precision and reliability of elasticity analysis in complicated burn scar cases. The need for a technique that would help physicians by objectively assessing elastic properties of scars, motivated our original algorithm. This algorithm successfully employed active contours for tracking and finite element models for strain analysis. However, the previous approach considered only one normal area and one abnormal area within the region of interest, and scar shapes which were somewhat simplified. Most burn scars have rather complicated shapes and may include multiple regions with different elastic properties. Hence, we need a method capable of adequately addressing these characteristics. The new method can split the region into more than two localities with different material properties, select and quantify abnormal areas, and apply different forces if it is necessary for a better shape description of the scar. The method also demonstrates the application of scale and mesh refinement techniques in this important domain. It is accomplished by increasing the number of Finite Element Method (FEM) areas as well as the number of elements within the area. The method is successfully applied to elastic materials and real burn scar cases. We demonstrate all of the proposed techniques and investigate the behavior of elasticity function in a 3-D space. Recovered properties of elastic materials are compared with those obtained by a conventional mechanics-based approach. Scar ratings achieved with the method are correlated against the judgments of physicians. [ABSTRACT FROM AUTHOR]

The progress of incorporating deep learning in the field of medical image interpretation has been greatly hindered due to the tremendous cost and time associated with generating ground truth for supervised machine learning, alongside concerns about the inconsistent quality of images acquired. Active learning offers a potential solution to these problems of expanding dataset ground truth by algorithmically choosing the most informative samples for ground truth labeling. Still, this effort incurs the costs of human labeling, which needs minimization. Furthermore, automatic labeling approaches employing active learning often exhibit overfitting tendencies while selecting samples closely aligned with the training set distribution and excluding out-of-distribution samples, which could potentially improve the model's effectiveness. We propose that the majority of out-of-distribution instances can be attributed to inconsistent cross images. Since the FDA approved the first whole-slide image system for medical diagnosis in 2017, whole-slide images have provided enriched critical information to advance the field of automated histopathology. Here, we exemplify the benefits of a novel deep learning strategy that utilizes high-resolution whole-slide microscopic images. We quantitatively assess and visually highlight the inconsistencies within the whole-slide image dataset employed in this study. Accordingly, we introduce a deep learning-based preprocessing algorithm designed to normalize unknown samples to the training set distribution, effectively mitigating the overfitting issue. Consequently, our approach significantly increases the amount of automatic region-of-interest ground truth labeling on high-resolution whole-slide images using active deep learning. We accept 92% of the automatic labels generated for our unlabeled data cohort, expanding the labeled dataset by 845%. Additionally, we demonstrate expert time savings of 96% relative to manual expert ground-truth labeling. [ABSTRACT FROM AUTHOR]

Studies a vision-based technique for objective assessment of burn scars. Application of active contours; Nonlinear finite element modeling of skin stretching; Details on burn scar assessment technique.

Presents information on a study which determined the dynamic-scale model construction from range imagery. Surface representations in a scale continuum; Surface model characteristics.

This paper describes a new algorithm which uses Gaussian and mean curvature values of the terrain surface to extract feature points. These points are then used for recognition of particular subregions of the terrain and in estimating relative positions of these subregions in the terrain. The Gaussian and mean curvatures are chosen because they are invariant under rotation and translation. In the Gaussian and mean curvature images, the points of maximum and minimum curvatures are extracted and used for matching. The stability of the position of these points in the presence of noise and with resampling is investigated. The input for this algorithm is 3D digital terrain data. Curvature values are calculated from the data by fitting a quadratic surface over a square window and calculating directional derivatives of this surface. A method of surface fitting that is invariant to sensor-centered coordinate system transformation is suggested and implemented. Real terrain data are used in our experiments. The algorithm is tested with and without the presence of noise, and its performance is described. [ABSTRACT FROM AUTHOR]

We propose an approach for characterizing structural heterogeneity of lung cancer nodules using Computed Tomography Texture Analysis (CTTA). Measures of heterogeneity were used to test the hypothesis that heterogeneity can be used as predictor of nodule malignancy and patient survival. To do this, we use the National Lung Screening Trial (NLST) dataset to determine if heterogeneity can represent differences between nodules in lung cancer and nodules in non-lung cancer patients. 253 participants are in the training set and 207 participants in the test set. To discriminate cancerous from non-cancerous nodules at the time of diagnosis, a combination of heterogeneity and radiomic features were evaluated to produce the best area under receiver operating characteristic curve (AUROC) of 0.85 and accuracy 81.64%. Second, we tested the hypothesis that heterogeneity can predict patient survival. We analyzed 40 patients diagnosed with lung adenocarcinoma (20 short-term and 20 long-term survival patients) using a leave-one-out cross validation approach for performance evaluation. A combination of heterogeneity features and radiomic features produce an AUROC of 0.9 and an accuracy of 85% to discriminate long- and short-term survivors. [ABSTRACT FROM AUTHOR]

Background: Current guidelines for lung cancer screening increased a positive scan threshold to a 6 mm longest diameter. We extracted radiomic features from baseline and follow‐up screens and performed size‐specific analyses to predict lung cancer incidence using three nodule size classes (<6 mm [small], 6‐16 mm [intermediate], and ≥16 mm [large]). Methods: We extracted 219 features from baseline (T0) nodules and 219 delta features which are the change from T0 to first follow‐up (T1). Nodules were identified for 160 incidence cases diagnosed with lung cancer at T1 or second follow‐up screen (T2) and for 307 nodule‐positive controls that had three consecutive positive screens not diagnosed as lung cancer. The cases and controls were split into training and test cohorts; classifier models were used to identify the most predictive features. Results: The final models revealed modest improvements for baseline and delta features when compared to only baseline features. The AUROCs for small‐ and intermediate‐sized nodules were 0.83 (95% CI 0.76‐0.90) and 0.76 (95% CI 0.71‐0.81) for baseline‐only radiomic features, respectively, and 0.84 (95% CI 0.77‐0.90) and 0.84 (95% CI 0.80‐0.88) for baseline and delta features, respectively. When intermediate and large nodules were combined, the AUROC for baseline‐only features was 0.80 (95% CI 0.76‐0.84) compared with 0.86 (95% CI 0.83‐0.89) for baseline and delta features. Conclusions: We found modest improvements in predicting lung cancer incidence by combining baseline and delta radiomics. Radiomics could be used to improve current size‐based screening guidelines. We demonstrated that combining delta radiomics with baseline radiomics generally improved the performance statistics to predict lung cancer incidence when compared to using only baseline radiomic features. We note inconsistent results in the performance statistics when we comparing overall models compared to models based on nodule size. As such, our findings suggest there is a trade‐off in terms of performance using nodule size‐specific models vs. an overall model. [ABSTRACT FROM AUTHOR]

Video capsule endoscopy (VCE) is an innovation that has revolutionized care within the field of gastroenterology, but the time needed to read the studies generated has often been cited as an area for improvement. With the aid of artificial intelligence, various fields have been able to improve the efficiency of their core processes by reducing the burden of irrelevant stimuli on their human elements. In this study, we have created and trained a convolutional neural network (CNN) capable of significantly reducing capsule endoscopy reading times by eliminating normal parts of the video while retaining abnormal ones. Our model, a variation of ResNet50, was able to reduce VCE video length by 47% on average and capture abnormal segments on VCE with 100% accuracy on three VCE videos as confirmed by the reading physician. The ability to successfully pre-process VCE footage as we have demonstrated will greatly increase the practicality of VCE technology without the expense of hundreds of hours of physician annotated videos. [ABSTRACT FROM AUTHOR]

Glioma is the most common type of primary malignant brain tumor. Accurate survival time prediction for glioma patients may positively impact treatment planning. In this paper, we develop an automatic survival time prediction tool for glioblastoma patients along with an effective solution to the limited availability of annotated medical imaging datasets. Ensembles of snapshots of three dimensional (3D) deep convolutional neural networks (CNN) are applied to Magnetic Resonance Image (MRI) data to predict survival time of high-grade glioma patients. Additionally, multi-sequence MRI images were used to enhance survival prediction performance. A novel way to leverage the potential of ensembles to overcome the limitation of labeled medical image availability is shown. This new classification method separates glioblastoma patients into long- and short-term survivors. The BraTS (Brain Tumor Image Segmentation) 2019 training dataset was used in this work. Each patient case consisted of three MRI sequences (T1CE, T2, and FLAIR). Our training set contained 163 cases while the test set included 46 cases. The best known prediction accuracy of 74% for this type of problem was achieved on the unseen test set. [ABSTRACT FROM AUTHOR]

Microglial cells mediate diverse homeostatic, inflammatory, and immune processes during normal development and in response to cytotoxic challenges. During these functional activities, microglial cells undergo distinct numerical and morphological changes in different tissue volumes in both rodent and human brains. However, it remains unclear how these cytostructural changes in microglia correlate with region-specific neurochemical functions. To better understand these relationships, neuroscientists need accurate, reproducible, and efficient methods for quantifying microglial cell number and morphologies in histological sections. To address this deficit, we developed a novel deep learning (DL)-based classification, stereology approach that links the appearance of Iba1 immunostained microglial cells at low magnification (20×) with the total number of cells in the same brain region based on unbiased stereology counts as ground truth. Once DL models are trained, total microglial cell numbers in specific regions of interest can be estimated and treatment groups predicted in a high-throughput manner (<1 min) using only low-power images from test cases, without the need for time and labor-intensive stereology counts or morphology ratings in test cases. Results for this DL-based automatic stereology approach on two datasets (total 39 mouse brains) showed >90% accuracy, 100% percent repeatability (Test-Retest) and 60× greater efficiency than manual stereology (<1 min vs. ∼ 60 min) using the same tissue sections. Ongoing and future work includes use of this DL-based approach to establish clear neurodegeneration profiles in age-related human neurological diseases and related animal models. • Deep learning model that links appearance of Iba1 microglial at low power with number of cells based on unbiased stereology. • Total microglial cell numbers can be estimated, and treatment groups predicted in a high-throughput manner (>1 min). • This approach uses only low-power images, no need for labor-intensive stereology counts or morphology ratings for test case. • Results for DL-based automatic stereology approach on two datasets (39 mouse brains) showed more than 90% accuracy. • 100% percent repeatability (Test-Retest) and 60x greater efficiency than manual using the same tissue sections. [ABSTRACT FROM AUTHOR]

Lung cancer is by far the leading cause of cancer death in the US. Recent studies have demonstrated the effectiveness of screening using low dose CT (LDCT) in reducing lung cancer related mortality. While lung nodules are detected with a high rate of sensitivity, this exam has a low specificity rate and it is still difficult to separate benign and malignant lesions. The ISBI 2018 Lung Nodule Malignancy Prediction Challenge, developed by a team from the Quantitative Imaging Network of the National Cancer Institute, was focused on the prediction of lung nodule malignancy from two sequential LDCT screening exams using automated (non-manual) algorithms. We curated a cohort of 100 subjects who participated in the National Lung Screening Trial and had established pathological diagnoses. Data from 30 subjects were randomly selected for training and the remaining was used for testing. Participants were evaluated based on the area under the receiver operating characteristic curve (AUC) of nodule-wise malignancy scores generated by their algorithms on the test set. The challenge had 17 participants, with 11 teams submitting reports with method description, mandated by the challenge rules. Participants used quantitative methods, resulting in a reporting test AUC ranging from 0.698 to 0.913. The top five contestants used deep learning approaches, reporting an AUC between 0.87 – 0.91. The team’s predictor did not achieve significant differences from each other nor from a volume change estimate (p =.05 with Bonferroni-Holm’s correction). [ABSTRACT FROM AUTHOR]

Gene-expression microarray datasets often consist of a limited number of samples with a large number of gene-expression measurements, usually on the order of thousands. Therefore, dimensionality reduction is critical prior to any classification task. In this work, the iterative feature perturbation method (IFP), an embedded gene selector, is introduced and applied to four microarray cancer datasets: colon cancer, leukemia, Moffitt colon cancer, and lung cancer. We compare results obtained by IFP to those of support vector machine-recursive feature elimination (SVM-RFE) and the t-test as a feature filter using a linear support vector machine as the base classifier. Analysis of the intersection of gene sets selected by the three methods across the four datasets was done. Additional experiments included an initial pre-selection of the top 200 genes based on their p values. IFP and SVM-RFE were then applied on the reduced feature sets. These results showed up to 3.32% average performance improvement for IFP across the four datasets. A statistical analysis (using the Friedman/Holm test) for both scenarios showed the highest accuracies came from the t-test as a filter on experiments without gene pre-selection. IFP and SVM-RFE had greater classification accuracy after gene pre-selection. Analysis showed the t-test is a good gene selector for microarray data. IFP and SVM-RFE showed performance improvement on a reduced by t-test dataset. The IFP approach resulted in comparable or superior average class accuracy when compared to SVM-RFE on three of the four datasets. The same or similar accuracies can be obtained with different sets of genes. [ABSTRACT FROM AUTHOR]

Support vector machines (SVMs) can be trained to be very accurate classifiers and have been used in many applications. However, the training time and, to a lesser extent, prediction time of SVMs on very large data sets can be very long. This paper presents a fast compression method to scale up SVMs to large data sets. A simple bit-reduction method is applied to reduce the cardinality of the data by weighting representative examples. We then develop SVMs trained on the weighted data. Experiments indicate that bit-reduction SVM produces a significant reduction in the time required for both training and prediction with minimum loss in accuracy. It is also shown to typically be more accurate than random sampling when the data are not overcompressed. [ABSTRACT FROM AUTHOR]

A novel thin line detection algorithm for use in low-altitude aerial vehicles is presented. This algorithm is able to detect thin obstacles such as cables, power lines, and wires. The system is intended to be used during urban search and rescue operations, capable of dealing with low-quality images, robust to image clutter, bad weather, and sensor artifacts. The detection process uses motion estimation at the pixel level, combined with edge detection, followed by a windowed Hough transform. The evidence of lines is tracked over time in the resulting parameter spaces using a dynamic line movement model. The algorithm's receiver operating characteristic curve (ROC) is shown, based on a multi-site dataset with 86 videos with 10160 wires spanning in 5576 frames. [ABSTRACT FROM AUTHOR]

Abstract: An ensemble of clustering solutions or partitions may be generated for a number of reasons. If the data set is very large, clustering may be done on tractable size disjoint subsets. The data may be distributed at different sites for which a distributed clustering solution with a final merging of partitions is a natural fit. In this paper, two new approaches to combining partitions, represented by sets of cluster centers, are introduced. The advantage of these approaches is that they provide a final partition of data that is comparable to the best existing approaches, yet scale to extremely large data sets. They can be 100,000 times faster while using much less memory. The new algorithms are compared against the best existing cluster ensemble merging approaches, clustering all the data at once and a clustering algorithm designed for very large data sets. The comparison is done for fuzzy and hard-k-means based clustering algorithms. It is shown that the centroid-based ensemble merging algorithms presented here generate partitions of quality comparable to the best label vector approach or clustering all the data at once, while providing very large speedups. [Copyright &y& Elsevier]

Abstract: This paper presents a new, automated image- and feature-based pipeline of analysis algorithms to elucidate and quantify a discovery regarding supersonic cracks made using large-scale molecular dynamics computations. The first computational confirmation of supersonic cracks was made in recent years, along with the discovery of a two-step, discrete process of increasing crack velocity through the nucleation of a transonic daughter crack and the later nucleation of a supersonic granddaughter crack. This discovery was facilitated by the work presented here. The algorithm pipeline includes modeling, recognition and motion analysis of both the front most crack tips and the more subtle secondary wavefronts from the slower ancestor cracks. The algorithms employed include line extraction from Canny edge maps, feature modeling based on physical properties, and subsequent tracking of primary and secondary wavefronts. The model embeds anticipated propagation properties (physics-based framework) and adapts to changes in the data for unexpected aspects (data-driven modeling). This process is completely automated; it runs in real time on three different 834-frame sequences using 40 250MHz processors. Results supporting the discovery of the two-step transition to supersonic crack propagation in bilayer materials are presented in terms of both feature tracking and velocity analysis. [Copyright &y& Elsevier]

This paper presents an active learning method which reduces the labeling effort of domain experts in multi-class classification problems. Active learning is applied in conjunction with support vector machines to recognize underwater zooplankton from higher-resolution, new generation SIPPER II images. Most previous work on active learning with support vector machines only deals with two class problems. In this paper, we propose an active learning approach "breaking ties" for multi-class support vector machines using the one-vs-one approach with a probability approximation. Experimental results indicate that our approach often requires significantly less labeled images to reach a given accuracy than the approach of labeling the least certain test example and random sampling. It can also be applied in batch mode resulting in an accuracy comparable to labeling one image at a time and retraining. [ABSTRACT FROM AUTHOR]

We present a methodology for correcting color images taken in practical indoor environments, such as laboratories, factories, and studios, that explicitly models illuminant location, surface reflectance and geometry, and camera responsivity. We explicitly model surfaces by taking our color images with corresponding registered three-dimensional (3-D) range images, which provide surface orientation and location information for every point in the scene. We automatically detect regions where color correction should not be applied, such as specularities, coarse texture regions, and jump edges. This correction results In objective color measures of the imaged surfaces. This kind of integrated, comprehensive system of color correction has not existed until now. i.e., it is the first of its kind In computer vision. We demonstrate results of applying this methodology to read Sages for applications In photorealistic rerendering, skin lesion detection, burn scar color measurement, and general color image enhancement. We also have tested the method under different lighting configurations and with three different range scanners. [ABSTRACT FROM AUTHOR]

We present a system to recognize underwater plankton images from the shadow Image particle profiling evaluation recorder (SIPPER). The challenge of the SIPPER image set is that many Images do not have clear contours. To address that, shape features that do not heavily depend on contour information were developed. A soft margin support vector machine (SVM) was used as the classifier. We developed a way to assign probability after multiclass SYM classification. Our approach achieved approximately 90% accuracy on a collection of plankton images. On another larger image set containing manually unidentifiable particles, it also provided 75.6% overall accuracy. The proposed approach was statistically significantly more accurate on the two data sets than a C4.S decision tree and a cascade correlation neural network. The single SYM significantly outperformed ensembles of decision trees created by bagging and random forests on the smaller data set and was slightly better on the other data set. The 15-feature subset produced by our feature selection approach provided slightly better accuracy than using all 29 features. Our probability model gave us a reasonable rejection curve on the larger data set. [ABSTRACT FROM AUTHOR]

This paper presents a gesture recognition system for visualization navigation. Scientists are interested in developing interactive settings for exploring large data sets in an intuitive environment. The input consists of registered 3-D data. A geometric method using Bezier curves is used for the trajectory analysis and classification of gestures. The hand gesture speed is incorporated into the algorithm to enable correct recognition from trajectories with variations in hand speed. The method is robust and reliable: correct hand identification rate is 99.9% (from 1641 frames), modes of hand movements are correct 95.6% of the time, recognition rate (given the right mode) is 97.9%. An application to gesture-controlled visualization of 3D bioinformatics data is also presented. [Copyright &y& Elsevier]

Presents a generic knowledge-guided segmentation and labelling approach that uses fuzzy clustering algorithms which is applicable to any set of multi feature images of the same location. Details of the overall structure of the approach; Experiment to collect a set of color images; Types of knowledge used in the approach; Construction and refinement of the knowledge base for the system.

Develops a hierarchical method which integrates local analysis and global analysis to recover dense depth values and nonrigid motion from a sequence of two-dimensional satellite cloud images without knowledge of point correspondences. Analysis of the several segmented small regions of the cloud images; Proposed recursive algorithm to integrate local analysis; Estimation of both the structure and the three-dimensional motion correspondence to subpixel accuracy.

Presents a study which proposed a methodology for calibrating multiple light source locations in three-dimensional form images. Methodology used; Results; Conclusions.

A knowledge-guided approach to automatic classification of Coastal Zone Color images off the West Florida Shelf is described. The approach is used to identify red tides on the West Florida Shelf, as well as areas with high concentration of dissolved organic matter such as a river plume found seasonally along the West Florida coast over the middle of the shelf. The Coastal Zone Color images are initially segmented by the unsupervised Multistage Random Sampling Fuzzy c-Means algorithm. Then, a knowledge-guided system is applied to the centroid values of resultant clusters to label case I, case II waters, a dilute river plume (“green river”), and red tide. The domain knowledge base contains information on cluster distribution in feature space, as well as spatial information such as bathymetry data. Our knowledge base consists of a rule-guided system and an embedded neural network. From 60 images, after training the system, this procedure recognizes all 15 images which contained a river plume and 45 images without. The system can correctly classify 74% of the pixels that belong to the river plume, which provides a substantial advantage to users looking for offshore extensions of riverine influence. Red tides are also successfully identified in a time series of images for which ground truth confirmed the presence of a harmful bloom. [ABSTRACT FROM AUTHOR]

• Quantifying stained cells in a biological tissue is critical for basic research. • Deep Learning superior to handcrafted algorithms for automatic cell quantification. • Majority of current automatic methods for single-stained tissue sections. • An adaptive method is proposed for stain separation in multi-stained images. • Single-stain methods can be applied to multi-stained images after stain separation. Quantifying cells in a defined region of biological tissue is critical for many clinical and preclinical studies, especially in the fields of pathology, toxicology, cancer and behavior. As part of a program to develop accurate, precise and more efficient automatic approaches for quantifying morphometric changes in biological tissue, we have shown that both deep learning-based and hand-crafted algorithms can estimate the total number of histologically stained cells at their maximal profile of focus in Extended Depth of Field (EDF) images. Deep learning-based approaches show accuracy comparable to manual counts on EDF images but significant enhancement in reproducibility, throughput efficiency and reduced error from human factors. However, a majority of the automated counts are designed for single-immunostained tissue sections. To expand the automatic counting methods to more complex dual-staining protocols, we developed an adaptive method to separate stain color channels on images from tissue sections stained by a primary immunostain with secondary counterstain. The proposed method overcomes the limitations of the state-of-the-art stain-separation methods, like the requirement of pure stain color basis as a prerequisite or stain color basis learning on each image. Experimental results are presented for automatic counts using deep learning-based and hand-crafted algorithms for sections immunostained for neurons (Neu-N) or microglial cells (Iba-1) with cresyl violet counterstain. Our findings show more accurate counts by deep learning methods compared to the handcrafted method. Thus, stain-separated images can function as input for automatic deep learning-based quantification methods designed for single-stained tissue sections. [ABSTRACT FROM AUTHOR]

Describes a system which provides an automatic segmentation and labeling of tumors, in magnetic resonance images (MRI) of the human brain. Role of MRI data in tracking the size of brain tumors; Classification of all slices processed by the tumor segmentation system; Identification of the five primary steps of the segmentation system.

• We propose an approach to classify ASD risk, based on gaze and demographic information such as age and gender. • We provide an analysis of eye gaze patterns, validating the use of these features. • We propose the use of 4 different machine learning-based classification algorithms for helping to classify ASD. • We present an analysis of which features are strongest for classification of ASD risk. People with autism spectrum disorder (ASD) display impairments in social interaction and communication skills, as well as restricted interests and repetitive behaviors, which greatly affect daily life functioning. Current identification of ASD involves a lengthy process that requires an experienced clinician to assess multiple domains of functioning. Considering this, we propose a method for classifying multiple levels of risk of ASD using eye gaze and demographic feature descriptors such as a subject's age and gender. We construct feature descriptors that incorporate the subject's age and gender, as well as features based on eye gaze patterns. We also present an analysis of eye gaze patterns validating the use of the selected hand-crafted features. We assess the efficacy of our descriptors to classify ASD on a National Database for Autism Research dataset, using multiple classifiers including a random forest, C4.5 decision tree, PART, and a deep feedforward neural network. [ABSTRACT FROM AUTHOR]

Purpose: To study the variability in volume change estimates of pulmonary nodules due to segmentation approaches used across several algorithms and to evaluate these effects on the ability to predict nodule malignancy. Methods: We obtained 100 patient image datasets from the National Lung Screening Trial (NLST) that had a nodule detected on each of two consecutive low dose computed tomography (LDCT) scans, with an equal proportion of malignant and benign cases (50 malignant, 50 benign). Information about the nodule location for the cases was provided by a screen capture with a bounding box and its axial location was indicated. Five participating quantitative imaging network (QIN) institutions performed nodule segmentation using their preferred semi‐automated algorithms with no manual correction; teams were allowed to provide additional manually corrected segmentations (analyzed separately). The teams were asked to provide segmentation masks for each nodule at both time points. From these masks, the volume was estimated for the nodule at each time point; the change in volume (absolute and percent change) across time points was estimated as well. We used the concordance correlation coefficient (CCC) to compare the similarity of computed nodule volumes (absolute and percent change) across algorithms. We used Logistic regression model on the change in volume (absolute change and percent change) of the nodules to predict the malignancy status, the area under the receiver operating characteristic curve (AUROC) and confidence intervals were reported. Because the size of nodules was expected to have a substantial effect on segmentation variability, analysis of change in volumes was stratified by lesion size, where lesions were grouped into those with a longest diameter of <8 mm and those with longest diameter ≥ 8 mm. Results: We find that segmentation of the nodules shows substantial variability across algorithms, with the CCC ranging from 0.56 to 0.95 for change in volume (percent change in volume range was [0.15 to 0.86]) across the nodules. When examining nodules based on their longest diameter, we find the CCC had higher values for large nodules with a range of [0.54 to 0.93] among the algorithms, while percent change in volume was [0.3 to 0.95]. Compared to that of smaller nodules which had a range of [−0.0038 to 0.69] and percent change in volume was [−0.039 to 0.92]. The malignancy prediction results showed fairly consistent results across the institutions, the AUC using change in volume ranged from 0.65 to 0.89 (Percent change in volume was 0.64 to 0.86) for entire nodule range. Prediction improves for large nodule range (≥ 8 mm) with AUC range 0.75 to 0.90 (percent change in volume was 0.74 to 0.92). Compared to smaller nodule range (<8 mm) with AUC range 0.57 to 0.78 (percent change in volume was 0.59 to 0.77). Conclusions: We find there is a fairly high concordance in the size measurements for larger nodules (≥8 mm) than the lower sizes (<8 mm) across algorithms. We find the change in nodule volume (absolute and percent change) were consistent predictors of malignancy across institutions, despite using different segmentation algorithms. Using volume change estimates without corrections shows slightly lower predictability (for two teams). [ABSTRACT FROM AUTHOR]

A novel stereology approach, the automatic optical fractionator, is presented for obtaining unbiased and efficient estimates of the number of cells in tissue sections. Used in combination with existing segmentation algorithms and ordinary immunostaining methods, automatic estimates of cell number are obtainable from extended depth of field images built from three-dimensional volumes of tissue (disector stacks). The automatic optical fractionator is more accurate, 100% objective and 8–10 times faster than the manual optical fractionator. An example of the automatic fractionator is provided for counts of immunostained neurons in neocortex of a genetically modified mouse model of neurodegeneration. Evidence is presented for the often overlooked prerequisite that accurate counting by the optical fractionator requires a thin focal plane generated by a high optical resolution lens. [ABSTRACT FROM AUTHOR]

Purpose Many radiomics features were originally developed for non-medical imaging applications and therefore original assumptions may need to be reexamined. In this study, we investigated the impact of slice thickness and pixel spacing (or pixel size) on radiomics features extracted from Computed Tomography ( CT) phantom images acquired with different scanners as well as different acquisition and reconstruction parameters. The dependence of CT texture features on gray-level discretization was also evaluated. Methods and materials A texture phantom composed of 10 different cartridges of different materials was scanned on eight different CT scanners from three different manufacturers. The images were reconstructed for various slice thicknesses. For each slice thickness, the reconstruction Field Of View ( FOV) was varied to render pixel sizes ranging from 0.39 to 0.98 mm. A fixed spherical region of interest ( ROI) was contoured on the images of the shredded rubber cartridge and the 3D printed, 20% fill, acrylonitrile butadiene styrene plastic cartridge ( ABS20) for all phantom imaging sets. Radiomic features were extracted from the ROIs using an in-house program. Features categories were: shape (10), intensity (16), GLCM (24), GLZSM (11), GLRLM (11), and NGTDM (5), fractal dimensions (8) and first-order wavelets (128), for a total of 213 features. Voxel-size resampling was performed to investigate the usefulness of extracting features using a suitably chosen voxel size. Acquired phantom image sets were resampled to a voxel size of 1 × 1 × 2 mm3 using linear interpolation. Image features were therefore extracted from resampled and original datasets and the absolute value of the percent coefficient of variation (% COV) for each feature was calculated. Based on the % COV values, features were classified in 3 groups: (1) features with large variations before and after resampling (% COV >50); (2) features with diminished variation (% COV <30) after resampling; and (3) features that had originally moderate variation (% COV <50%) and were negligibly affected by resampling. Group 2 features were further studied by modifying feature definitions to include voxel size. Original and voxel-size normalized features were used for interscanner comparisons. A subsequent analysis investigated feature dependency on gray-level discretization by extracting 51 texture features from ROIs from each of the 10 different phantom cartridges using 16, 32, 64, 128, and 256 gray levels. Results Out of the 213 features extracted, 150 were reproducible across voxel sizes, 42 improved significantly (% COV <30, Group 2) after resampling, and 21 had large variations before and after resampling (Group 1). Ten features improved significantly after definition modification effectively removed their voxel-size dependency. Interscanner comparison indicated that feature variability among scanners nearly vanished for 8 of these 10 features. Furthermore, 17 out of 51 texture features were found to be dependent on the number of gray levels. These features were redefined to include the number of gray levels which greatly reduced this dependency. Conclusion Voxel-size resampling is an appropriate pre-processing step for image datasets acquired with variable voxel sizes to obtain more reproducible CT features. We found that some of the radiomics features were voxel size and gray-level discretization-dependent. The introduction of normalizing factors in their definitions greatly reduced or removed these dependencies. [ABSTRACT FROM AUTHOR]

Purpose Radiomics utilizes a large number of image-derived features for quantifying tumor characteristics that can in turn be correlated with response and prognosis. Unfortunately, extraction and analysis of such image-based features is subject to measurement variability and bias. The challenge for radiomics is particularly acute in Positron Emission Tomography (PET) where limited resolution, a high noise component related to the limited stochastic nature of the raw data, and the wide variety of reconstruction options confound quantitative feature metrics. Extracted feature quality is also affected by tumor segmentation methods used to define regions over which to calculate features, making it challenging to produce consistent radiomics analysis results across multiple institutions that use different segmentation algorithms in their PET image analysis. Understanding each element contributing to these inconsistencies in quantitative image feature and metric generation is paramount for ultimate utilization of these methods in multi-institutional trials and clinical oncology decision making. Methods To assess segmentation quality and consistency at the multi-institutional level, we conducted a study of seven institutional members of the National Cancer Institute Quantitative Imaging Network. For the study, members were asked to segment a common set of phantom PET scans acquired over a range of imaging conditions as well as a second set of head and neck cancer (HNC) PET scans. Segmentations were generated at each institution using their preferred approach. In addition, participants were asked to repeat segmentations with a time interval between initial and repeat segmentation. This procedure resulted in overall 806 phantom insert and 641 lesion segmentations. Subsequently, the volume was computed from the segmentations and compared to the corresponding reference volume by means of statistical analysis. Results On the two test sets (phantom and HNC PET scans), the performance of the seven segmentation approaches was as follows. On the phantom test set, the mean relative volume errors ranged from 29.9 to 87.8% of the ground truth reference volumes, and the repeat difference for each institution ranged between −36.4 to 39.9%. On the HNC test set, the mean relative volume error ranged between −50.5 to 701.5%, and the repeat difference for each institution ranged between −37.7 to 31.5%. In addition, performance measures per phantom insert/lesion size categories are given in the paper. On phantom data, regression analysis resulted in coefficient of variation (CV) components of 42.5% for scanners, 26.8% for institutional approaches, 21.1% for repeated segmentations, 14.3% for relative contrasts, 5.3% for count statistics (acquisition times), and 0.0% for repeated scans. Analysis showed that the CV components for approaches and repeated segmentations were significantly larger on the HNC test set with increases by 112.7% and 102.4%, respectively. Conclusion Analysis results underline the importance of PET scanner reconstruction harmonization and imaging protocol standardization for quantification of lesion volumes. In addition, to enable a distributed multi-site analysis of FDG PET images, harmonization of analysis approaches and operator training in combination with highly automated segmentation methods seems to be advisable. Future work will focus on quantifying the impact of segmentation variation on radiomics system performance. [ABSTRACT FROM AUTHOR]