Issue 41Issue • April24 • 2017
TABLE OF CONTENTS RECENT RELEASES
Editor's Note...... 1 KITWARE AND COLLABORATORS MAKE 1.0 RELEASE OF MATERIALS TOMOGRAPHY Recent Releases...... 1 PLATFORM Kitware worked with researchers to make the first major Novel Mechanism Segments Anatomical Structures for 3D release of Tomviz, an open-source platform that processes Printing...... 3 data for materials tomography. Tomviz aligns, reconstructs, segments, visualizes, and analyzes three-dimensional tomo- GeoJS Draws Better Lines...... 7 graphic data through a reproducible pipeline. Kitware News...... 9 The release followed the conclusion of Phase II of a Department of Energy Small Business Innovation Research project. For the project, Kitware partnered with Cornell EDITOR'S NOTE University and collaborated with members of other academic institutions to develop a multithreaded application architec- As winter turned to spring, we experienced some changes ture and an extensive Python environment in the platform. of our own at Kitware. For starters, we re-designed our “We created a responsive, graphical solution that processes website. The re-design not only consolidated content and and analyzes complex scientific data in a manner that may navigation options, but it revitalized the look of our pages. otherwise span multiple applications and programming Our website now runs on WordPress, scales better on cell environments,” said Marcus Hanwell, a technical leader at phones and tablets, and provides additional visualizations Kitware. Hanwell is a member of the Tomviz development and information on areas of focus. Please have a look at team who served as the principal investigator on the project. https://www.kitware.com. Let us know what you think!
Since January, we also grew our team; we saw long-time employees jump into new roles; and we moved our office location in Santa Fe, New Mexico. To aid our communities, we collected pet supplies such as toys, food, and blankets. In addition, we sold paper hearts for Ronald McDonald House Charities. Our Kitware Gives Back team is already working on upcoming initiatives. For specifics, please read our blog at https://blog.kitware.com.
This year, we are a silver sponsor of Computer Vision and Pattern Recognition (CVPR). Members of our computer vision team will set up a booth (242) at the event in July 2017. Multiple team members are event chairs. Senior Director of Computer Vision Anthony Hoogs is a general chair, and Assistant Director of Computer Vision Matt Turek is a corporate relations chair. We have more details in an event listing at https://blog.kitware.com/events/computer- vision-and-pattern-recognition-cvpr-2017.
Tomviz shows an animation of a nanotube from Robert Hovden. Tomviz comes ready for download as open-source binaries for CMAKE UPS SUPPORT FOR CUDA, CSHARP, major operating systems like Windows, macOS, and Linux. To AND C++ begin the pipeline, Tomviz imports projection images from Kitware released version 3.8 of its open-source solution for transmission electron microscopes in industry-standard file cross-platform software development, CMake. The release formats such as Tagged Image File Format (TIFF), Medical upgraded support for CUDA, CSharp (C#), and C++, promot- Research Council (MRC), and Electron Microscopy Datasets ing both CUDA and C# to first-class programming languages. (EMD). While the platform can automatically align these Likewise, the release further simplified project compilation images, it enables users to manually adjust the tilt and rota- with the addition of C++11, C++14, and C++17 target-based tion of the axis. feature flags. These flags extend the existingtarget_ command. Kitware published detailed “Throughout the pipeline, Tomviz remains highly interac- compile_features notes on CMake 3.8 on its blog at https://blog.kitware.com/ tive,” said Robert Hovden, a former postdoctoral researcher cmake-3-8-0-available-for-download. To download CMake, at Cornell University and a founder of the Tomviz develop- please visit https://cmake.org/download. ment team. “Tomviz runs data operators in background threads, so users can continue to work with the platform PARAVIEW GETS MORE ANIMATED while it executes operations.” The ParaView 5.3 release included new plug-ins. One plug- in, , adjusts the gradient, contours, and details The platform builds on the capabilities of the Insight VTKmFilters in visualizations on many-core systems. Another plug-in, Segmentation and Registration Toolkit, the Visualization , animates streamlines. Toolkit, and ParaView. Through hardware-accelerated StreamLinesRepresentation rendering techniques, for example, Tomviz lets users alter visualization parameters such as the opacity, lighting, color, and orientation of volumes and geometries. Tomviz not only provides different rendering modes, but it contains multiple tools for analysis. One tool, a ruler, measures the lengths of objects in images. Tomviz also computes surface area, volume, and other labeled object attributes, which it displays in a spreadsheet that users can save.
In addition to tabular data, Tomviz allows users to save its state with all data pipeline settings. Users can share and modify these settings to help others follow the necessary steps to reproduce results. Tomviz saves results as movies or exports them on the web as HTML files. To foster open dissemination, the Tomviz development team made live A screenshot captures the use of the examples publicly available on GitHub. These examples StreamLinesRepresentation plug-in. utilize images or the latest Web Graphics Library (WebGL) To reduce aliasing artifacts when ParaView renders lines, rendering techniques. wireframes, etc., the release turned on the Fast Approximate This material is based upon work supported by the Anti-Aliasing (FXAA) algorithm by default. The release also Department of Energy under Office of Science Award updated the integration of ParaView and OSPRay through Number DE-SC0011385. This report was prepared as an support for OSPRay 1.1.2. For more information, please visit account of work sponsored by an agency of the United States https://blog.kitware.com/paraview-5-3-0-release-notes. Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any war- GIRDER 2.2 TUNES DATA MANAGEMENT Girder 2.2 became available as open-source software. The ranty, express or implied, or assumes any legal liability or release polished the user interface on the administrative responsibility for the accuracy, completeness, or usefulness configuration page. It also presented a file-like interface for of any information, apparatus, product, or process disclosed, files in Girder assetstores. In addition, the release augmented or represents that its use would not infringe privately owned support for file formats in the DICOM viewer plug-in, it built rights. Reference herein to any specific commercial product, in support for Box as an Open Authorization (OAuth) pro- process, or service by trade name, trademark, manufacturer, vider, it incorporated support for static-only dependencies or otherwise does not necessarily constitute or imply its between plug-ins for web client code, it introduced a mode endorsement, recommendation, or favoring by the United that boosts upload performance, and it refined download States Government or any agency thereof. The views and performance in the Secure File Transfer Protocol (SFTP) opinions of authors expressed herein do not necessarily server. To download Girder, please go to https://github.com/ state or reflect those of the United States Government or girder/girder/releases/tag/v2.2.0. any agency thereof.
2 NOVEL MECHANISM SEGMENTS ANATOMICAL STRUCTURES FOR 3D PRINTING
Csaba Pinter, Andras Lasso, and Gabor Fichtinger (Queen's University)
A fundamental task in most aspects of medical image Unfortunately, representing and processing anatomical computing is segmentation, i.e., delineation of anatomical structures present major difficulties. These difficulties include structures of interest for further processing and quantifica- operation, as it may be necessary to perform conversion. tion. Segmentation can be manual, it can be semi-automatic Another difficulty is identity, since it is important to keep (through the initialization of an algorithm with limited track of the origin (provenance) of the structures and what input), or it can be fully automatic (through an autonomous they represent. A third difficulty is validity: Representations algorithm). A multitude of software tools and algorithms may change after conversions, so invalid data must remain exist for each type of segmentation, and segmentation has inaccessible at all times. Coherence serves as an additional served as the subject of extensive research in the field of point of difficulty. As structure sets typically correspond to medical image computing. the same entity (i.e., the patient), the in-memory objects that relate to the structures need to form a unified whole. Most commonly, three-dimensional (3D) binary volumes (labelmaps) store segmentation results. Each volume simply 3D Slicer [2] is one of the most popular open-source platforms indicates whether a volumetric element (voxel) is inside in the world for medical image analysis and visualization. The or outside of the structure of interest. Three-dimensional Laboratory for Percutaneous Surgery (PerkLab) at Queen's binary volumes are optimal for most processing algorithms. University uses 3D Slicer extensively for a wide variety of To visualize structures, however, surface models are optimal. research projects. Previously, representing segmentations in Instead of a structured grid of voxels, each surface model 3D Slicer posed considerable hurdles, not only in terms of consists of a point cloud. Triangles connect the point cloud, the four above mentioned difficulties but also in regards to which a 3D visualization can illustrate. processing time. Researchers and clinicians spent time manu- ally converting one representation to another. For example, Several other representations exist. In radiation therapy they used the module to create 3D surfaces of (RT), the Digital Imaging and Communications in Medicine ModelMaker labelmaps. They did this every time the labelmaps changed. (DICOM) standard [1] specifies planar contours. In addition, certain segmentation algorithms yield labelmaps with voxels A MECHANISM FOR DYNAMIC ANATOMICAL that indicate probabilities instead of binary decisions. Even STRUCTURE MANAGEMENT more obscure representations include ribbons, which help to PerkLab designed and implemented a new data type in quickly visualize planar contours. 3D Slicer (and before that in the SlicerRT extension [3]), called segmentation node. This data type contains multiple structures and multiple representations in the same object, which addresses the “identity” difficulty. (See the follow- ing image.) The structures can remain synchronized after changes occur in the underlying data, which attends to the “coherence” difficulty.
The data type is part of a complete mechanism that manages the contained representations. Whenever 3D Slicer requests a representation—such as when the 3D view tries to access the surface model of a segmentation—the mechanism automatically performs conversion. This overcomes the “operation” difficulty.
A master representation makes sure that invalid data does not become available, which tackles the “validity” difficulty. The master representation is the representation in which the data creation initially occurred (i.e., the labelmap, when segmenting manually or semi-automatically, and the planar A brain stem structure appears in different representations. contour, when loading from DICOM-RT). The master repre- "A" shows planar contours, "B" shows a closed surface, "C" sentation also serves as the source of all conversions. When shows a binary labelmap, and "D" shows a ribbon model. the master representation changes, e.g., during the segmen-
3 tation of an organ, all other representations get cleared. module, which served as the starting point for the Segment The representations get re-converted, when they become Editor design. The user interface of Segment Editor is requested again. very similar to that of Editor, which makes the transition easier for users. The underlying implementation, however, is very different. It is, in some ways, simpler. New features include these:
• real-time visualization of the 3D surface that results from the current state of the edited labelmap; • support for overlapping segments (e.g., structures) and advanced masking options;
• a Segments Table panel, which allows robust per- structure handling and provides advanced visualization settings; • brushes that paint in 3D rather than in a slice-by-slice manner;
• a Segment Editor widget, which any module or slice- let (i.e., a stripped-down user interface that focuses on a single task and provides a streamlined workflow) can embed; • editing options for oblique slices that do not align with the main axes of the anatomical volume; and
• new editor tools or effects such as Scissors, Fill between slices, Grow from seeds, etc. The overall mechanism fills the basic needs of manual and semi-automatic segmentation. Thus, a wide range of appli- cations can benefit from it. Since the mechanism makes it easy to convert representations from one data type to segmentation node, users can rely on the mechanism, even An example of the segmentation node data type if their data comes from a third-party tool. stores each structure of an entity (patient) with each representation in one place. The aforementioned field of RT was the first target area for the mechanism, as most workflows in the field use numerous PerkLab carefully chose and implemented conversion algo- types of data representations. Visualizing and analyzing RT rithms to cover the widest variety of datasets. (For more datasets became immensely more straightforward and robust information, see "Reconstruction of surfaces from planar since the adoption of the mechanism. More specialized use contours through contour interpolation" [4].) A directed graph cases have also benefited from its utilization. These use cases drives the algorithms. The graph contains the representa- include the fusion of magnetic resonance and ultrasound tions as nodes and the conversion algorithms as edges. It images for brachytherapy applications [6]. Use cases also attempts to find the computationally cheapest conversion include the evaluation of dosimetric measurements with gel between the master representation and the requested rep- [7] and film dosimeters. To facilitate more widespread adop- resentation. tion, however, a more prominent and mainstream use case needed to demonstrate and prove the mechanism. APPLICATIONS The infrastructure of the mechanism potentially spans every 3D PRINTING TUTORIAL workflow that includes segmented anatomical structures. As 3D printing is a popular and continually growing field, and the core implementation only depends on the Visualization so it offers a natural use case that presents the power of the Toolkit (VTK) [5], a wide variety of software tools in the field new module and infrastructure. A recent tutorial [8] walks of medical image computing are already compatible for through the necessary steps to create a spine phantom from adoption. Similarly, in 3D Slicer, the infrastructure poten- a partial spine structure and a phantom base. In the tutorial, tially spans all segmentation-related operations. 3D Slicer segments the partial spine structure on a computed tomography (CT) image. The phantom base comes from a The design and implementation of the new Segment separate computer-aided design (CAD) application. The base module started from scratch. The module aspires Editor facilitates reproducible and stable placement of the phantom to be the main module for manual and semi-automatic in its plastic container, and it incorporates a marker holder segmentation. It is the successor of the 3D Slicer Editor into which electromagnetic (EM) sensors can accurately fit.
4 Thus, the tutorial exemplifies the use of an external CAD sions. Using the above four effects, however, it is possible design with a custom segmentation from 3D Slicer. to create a printable patient-specific phantom, as it is easy to export segmentation results to a STereoLithography (STL) A simulation for facet joint injection training [9] inspired the file that is ready for 3D printing. design of the phantom. The simulation teaches trainees the complex technique of needle insertion but does not involve a SUMMARY live patient. It aims to support competency-based education To address a very common need in medical image comput- by quantitatively evaluating the collected electromagnetic ing, PerkLab implemented a complex software infrastructure tracking and video data with the PerkTutor extension in 3D Slicer that facilitates more automated and structured [10]. The extension is a training platform for image-guided handling of segmentation results. PerkLab created Segment interventions. Editor to provide an easy-to-use tool to manually or semi- automatically segment anatomical structures.
3D Slicer 4.7 creates the final phantom model.
A tutorial demonstrates the features of these developments. The tutorial results in a printed training phantom. During the tutorial, 3D Slicer generates a partial spinal The tutorial uses a recent 3D Slicer nightly release instead column segmentation and merges it with a CAD model to of version 4.6.2. Although the modules in version 4.6.2 are create a training phantom that can become printed in three stable, the 3D Slicer community continues to add useful fea- dimensions. tures to Segment Editor. These features increase usability, DISCUSSION and they provide new and convenient segmentation effects. This mechanism and related modules have undergone devel- The tutorial can run on all of the operating systems that opment for years, over multiple iterations, and they are now support 3D Slicer, which include Windows, macOS, and stable. In the spirit of continual improvement, however, Linux. The featured effects are as follows. PerkLab encourages the community to contact developers to convey their ideas and comments via the 3D Slicer forum • : This effect provides a very simple seg- Threshold on Discourse [11]. mentation tool that creates a segment from a range of intensity values in the anatomical image. It also specifies an To further help the community, PerkLab plans to create a intensity-based mask. video tutorial of the same workflow with more examples that use . Also, PerkLab plans to add more • Islands: This effect handles connected components in a Segment Editor segment. modules to the mechanism such as the Segment Statistics module, which calculates volume, image intensity, and other • Scissors: This effect selects areas in two dimensions and statistics on segmented structures. Other work will integrate volumes in three dimensions. Segment Editor can erase these selections, or it can fill inside or outside of them. fractional labelmaps. Although these store segmentations in the same resolution as binary labelmaps, the segmentations • : This effect contains a set of mor- Logical Operators have considerably higher accuracy. In addition, fractional phological operators that add, subtract, intersect, etc. labelmaps provide support for probabilistic segmentations. Segment Editor includes many other editor effects, and Updates will follow. the module can obtain more through downloadable exten-
5 ACKNOWLEDGMENT Csaba Pinter is a research software engineer Funding for this work came in part from Cancer Care Ontario, at Queen’s University in Canada. He spent through Applied Cancer Research Unit and Research Chair several years in the industry, working on in Cancer Imaging grants, and from Ontario Consortium for projects for medical image computing. He Adaptive Invention in Radiation Oncology. now delves into the world of open source. He is an active contributor to 3D Slicer, and REFERENCES he is the owner of the SlicerRT toolkit for radiation 1. Mildenberger, Peter, Marco Eichelberg, and Eric Martin. "Introduction to the DICOM standard." European Radiology therapy. His main interest is the design of innovative medical 12 (2002): 920-927. applications. 2. Fedorov, Andriy, Reinhard Beichel, Jayashree Kalpathy- Cramer, Julien Finet, Jean-Christophe Fillion-Robin, Sonia Andras Lasso is a senior software engineer Pujol, Christian Bauer, Dominique Jennings, Fiona Fennessy, and the associate director of PerkLab. He Milan Sonka, John Buatti, Stephen Aylward, James V. Miller, joined PerkLab in 2009, after he spent nine Steve Pieper, and Ron Kikinis. "3D Slicer as an image com- years at GE Healthcare as a software engi- puting platform for the Quantitative Imaging Network." Magnetic Resonance Imaging 30 (2012): 1323-1341. neer. His main interests are the development 3. Pinter, Csaba, Andras Lasso, An Wang, David Jaffray, and of high-quality reusable open-source soft- Gabor Fichtinger. "SlicerRT: Radiation therapy research ware components and the employment of these components toolkit for 3D Slicer." Medical Physics 39 (2012): 6332-6338. for building systems for translational research and clinical 4. Sunderland, Kyle, Boyeong Woo, Csaba Pinter, and Gabor Fichtinger. "Reconstruction of surfaces from planar use. contours through contour interpolation." In Proc. SPIE, 94151R. Medical Imaging 2015: Image-Guided Procedures, Gabor Fichtinger is currently the director of Robotic Interventions, and Modeling, Orlando, Florida. The PerkLab. He received his doctoral degree in International Society for Optics and Photonics, 2015. computer science from Budapest University 5. Schroeder, Will, Ken Martin, and Bill Lorensen. The of Technology and Economics in 1990. He is a Visualization Toolkit: An Object-Oriented Approach to 3D Graphics (4th edition). Clifton Park: Kitware, 2006. professor of computer science, electrical 6. Poulin, Eric, Karim Boudam, Csaba Pinter, Samuel Kadoury, engineering, mechanical engineering, and Andras Lasso, Gabor Fichtinger, and Cynthia Ménard. surgery at Queen’s University, with adjunct appointments at "Validation of MRI To US Registration For Focal HDR Johns Hopkins University. He specializes in robot-assisted Prostate Brachytherapy Treatment." In proceedings of the 2017 ABS annual meeting. The Value of Brachytherapy image-guided needle-placement procedures, primarily for in Multidisciplinary Cancer Care, Boston, Massachusetts. cancer diagnosis and therapy. American Brachytherapy Society, 2017. 7. Alexander, Kevin M., Csaba Pinter, Jennifer Andrea, Gabor Fichtinger, and L. John Schreiner. (2015). "3D Slicer gel dosimetry analysis: Validation of the calibration process." In IFMBE Proceedings, 521-524. World Congress on Medical Physics and Biomedical Engineering, Toronto, Canada. Springer International Publishing, 2015. 8. Pinter, Csaba. "Segmentation for 3D Printing" (tutorial presented at the 24th project week of the National Alliance for Medical Image Computing, Cambridge, Massachusetts, January 9-13, 2017). http://www.na-mic. org/Wiki/images/a/a6/SegmentationFor3DPrinting_ TutorialContestWinter2017.pdf. 9. Moult, Eric, Tamas Ungi, Mattea Welch, J. Lu, Robert C. McGraw, and Gabor Fichtinger. "Ultrasound-guided facet joint injection training using Perk Tutor." International Journal of Computer Assisted Radiology and Surgery 8 (2013): 831-836. 10. Ungi, Tamas, Derek Sargent, Eric Moult, Andras Lasso, Csaba Pinter, Robert C. McGraw, and Gabor Fichtinger. "Perk Tutor: An Open-Source Training Platform for Ultrasound- Guided Needle Insertions." IEEE Transactions on Biomedical Engineering 59 (2012): 3475-3481. 11. "3D Slicer." April 19, 2017. https://discourse.slicer.org.
6 GEOJS DRAWS BETTER LINES
David Manthey, Aashish Chaudhary (Kitware)
Lines are a basic component of most data visualizations. popular open-source geographic information system librar- Lines can convey much more than just edges and connec- ies that show data in a geospatial context, none of them vary tivity; the color, thickness, and opacity of lines can provide properties along the length of a line. Furthermore, GeoJS additional information. GeoJS, an open-source library for out-performs these libraries in the number of lines that it geospatial and large-scale image visualization, can draw can draw at once. With its many features for line drawing, sophisticated lines. The recent version 0.11.0 release made GeoJS produces clearer and more informative visualizations. many stylistic and quality improvements to this ability. The following example uses data from historic hurricanes. GeoJS is a client-side web rendering library. This means that it does not have to communicate with any server. Rather, it runs entirely in a web browser, which allows local resources to draw maps or annotations and enables visualizations to scale. GeoJS supports line drawing via Scalable Vector Graphics (SVG),
SVG uses an eXtensible Markup Language (XML) specifica- tion. While SVG is a popular backend for drawing graphics, it has limitations in regards to rendering lines. The SVG render- ing process is very slow when it needs to draw many lines. In addition, SVG cannot vary line width or add color along the length of a line. Since a browser tracks each SVG item that gets drawn, it is easy to trigger events. The browser can show information, for example, when a mouse hovers over a line. SVG does not, however, scale to millions of items. In this case, the path of each hurricane is a simple line. The HyperText Markup Language (HTML)
Web Graphics Library (WebGL) exposes access to graph- ics acceleration and OpenGL. WebGL programs everything explicitly, which gives greater flexibility, but it takes longer to create basic features. WebGL requires either graphics pro- cessing unit (GPU) support or emulation. With WebGL, only the GPU limits the scale of the data.
Although most browsers support SVG,
7 Another example uses air traffic information from the OpenSky Network on http://www.opensky-network.org. This information includes aircraft location, ground speed, altitude, and other parameters.
Lines change color with smooth transitions through their lengths.
GeoJS correlates wind speed and pressure. It varies the width This image shows one hour of data. of each line to show wind speed. GeoJS plots altitude in color, where red represents the highest altitude, and blue represents the lowest altitude. In addition, GeoJS encodes speed in line width, as wider lines indicate faster speeds. Furthermore, GeoJS shows time with opacity. It makes older data more transparent.
As this image demonstrates, some hurricanes with high- speed winds do not have correspondingly extreme instances of low pressure. In a single view, GeoJS illustrates the direction of travel, the relative speed, and the altitude of various flights. GeoJS can The visualization of the hurricane tracks looks nicer with plot thousands of flights concurrently. rounded line segment joins. Line rendering is essential for visualizing information from geospatial datasets (e.g., flight paths and navigation routes) and climate datasets (e.g., hurricane tracks, contour lines, and even streamlines). The new features for line drawing in GeoJS enable the creation of sophisticated visualizations that can more readily and clearly convey information. GeoJS makes it possible, for example, to draw hundreds of thou- sand of line segments and still achieve good performance. Thus, it helps to easily explore and analyze large datasets. With support for common line drawing features such as join shape; miter limit, which determines how to join sharp lines; and antialiasing, which smooths lines to make crisp-looking edges, GeoJS allows for adjustments that best represent data and are aesthetically pleasing. In particular, the path that jogs to the north and then to the southwest looks good with rounded line segment joins.
8 ACKNOWLEDGEMENT The authors thank the Science Mission Directorate at NASA headquarters for providing this opportunity. KITWARE NEWS The material is based upon work supported by NASA under Award Number NNX15AE30G. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Aeronautics and Space Administration. PAPER OFFERS NEW APPROACHES TO David Manthey is a research and develop- APPLICATION DEVELOPMENT ment engineer for the high-performance Kitware introduced two approaches at IEEE Virtual Reality computing (HPC) and visualization team at (VR) 2017 that greatly simplify VR application development Kitware. He received a B.S. in mechanical through the use of the Visualization Toolkit (VTK). The engineering from Rensselaer Polytechnic introduction occurred during the presentation of a paper Institute in 1992. He has worked in computer that members of Kitware co-authored with researchers from vision, measuring three-dimensional geometry and defor- Indiana University and Idaho National Laboratory. mation in sheet metal and other materials. More recently, he “Previously, application developers took a significant amount has worked on audio and video distribution, storage, and of time to create scientific visualization VR applications, as processing. These efforts range from direct hardware control they had to write their own algorithms and data-reading to user interfaces. He has one patent that relates to the routines,” said Patrick O’Leary, a co-author on the paper automatic calibration of internal and external camera and an assistant director of scientific computing at Kitware. parameters. “With VTK, developers can access ready-to-use algorithms Aashish Chaudhary is technical leader for through a convenient application programming interface. the HPC and visualization team at Kitware. “These algorithms not only save developers time, but they As an active researcher and developer in offer unmatched capabilities for rendering and interacting information and scientific visualization, he with data.” has contributed in various technical capaci- ties to the Visualization Toolkit (VTK), In the paper, “Enhancements to VTK Enabling Scientific ParaView, Ultrascale Visualization-Climate Data Analysis Visualization in Immersive Environments,” O’Leary and co- Tools (UV-CDAT), the VTK OpenGL ES Rendering Toolkit authors describe OpenGL context sharing and VR toolkit (VES), and many other Small Business Innovation Research embedding. Respectively, these approaches allow develop- and open-source projects. ers to use VTK with consumer-level devices such as Oculus Rift and HTC Vive as well as with immersive environments including CAVE™ systems.
Kitware officially added VR support to VTK in version 7.1 with the integration of two modules. Along with VTK, the modules equip ParaView with VR functionality. ParaView is an open-source software solution that analyzes large datasets and employs VTK as its rendering engine. Kitware cited the inclusion of VR functionality in ParaView as one of the notable achievements that led to recent recogni- tion. In February 2017, the company received the Software Innovation award from Albany Business Review, and in November 2016, Kitware accepted an award from HPCwire for Editor’s Choice: Best HPC Visualization Product or Technology on behalf of the ParaView development com- munity.
In future efforts, Kitware plans to bring support for portable, threaded data parallel algorithms to VTK.
“We aim to continue to extend VTK and ParaView to help developers explore data so that they can more easily make scientific discoveries,” O’Leary said.
Two of O’Leary’s co-authors on the IEEE VR 2017 paper, Sankhesh Jhaveri, a senior research and development
9 engineer at Kitware, and William Sherman, a senior techni- guidance in designing and implementing advanced algo- cal advisor and team lead at Indiana University, presented rithms for the many open-source platforms that Kitware their work on VTK and ParaView during the Systems and develops. Applications session at the conference. The session took “At Kitware, we produce software that scientists and engi- place Wednesday, March 22, 2017, from 1:30 to 3 p.m. PST, in neers use to change the world every day,” Schroeder said. Los Angeles, California. “We have grown significantly over the past two decades, This material is based upon work supported by the building a vibrant culture with a collaborative focus that Department of Energy under Award Number DE-SC0010119. promotes technical excellence.” This report was prepared as an account of work sponsored At the shareholders meeting, Kitware co-founder and current by an agency of the United States Government. Neither the Chief Technical Officer Bill Hoffman accepted the responsi- United States Government nor any agency thereof, nor any bilities of chairman of the board. As chairman, Hoffman will of their employees, makes any warranty, express or implied, work closely with Avila to guide company growth. or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, Kitware appointed Claudine Hagen to succeed Kitware co- apparatus, product, or process disclosed, or represents that founder Ken Martin, Ph.D., as chief financial officer. Hagen its use would not infringe privately owned rights. Reference joined the company as director of finance in January 2013. herein to any specific commercial product, process, or service In this role, she led and coordinated business planning, by trade name, trademark, manufacturer, or otherwise does accounting, financial reporting, and budgeting efforts. not necessarily constitute or imply its endorsement, recom- Prior to joining Kitware, Hagen spent 13 years in public mendation, or favoring by the United States Government accounting with KPMG, where she gained extensive experi- or any agency thereof. The views and opinions of authors ence partnering with senior management teams to develop expressed herein do not necessarily state or reflect those of business strategies and to optimize accounting and finance the United States Government or any agency thereof. functions.
Research reported in this publication was supported by the For his new title of distinguished engineer, Martin will direct National Institute Of Biomedical Imaging And Bioengineering technology initiatives that pertain to rendering, virtual of the National Institutes of Health under Award Number reality, and data visualization. Kitware co-founder Charles R01EB014955. The content is solely the responsibility of the Law, Ph.D., who performed the duties of vice president of authors and does not necessarily represent the official views strategic growth, also became a distinguished engineer. He of the National Institutes of Health. will help introduce Kitware to new markets that closely align with its areas of expertise in high-performance computing LEADERSHIP TRANSITIONS FOSTER GROWTH (HPC) and visualization, data and analytics, medical imaging, Less than three weeks after it celebrated its 19th anniversary, computer vision, and quality software process. Kitware made transitions in team management and orga- nizational structure to foster growth and to strengthen its Brad Davis, Ph.D., who has worked at Kitware for more than position as a leader in open-source technical computing. The 10 years in a variety of technical and business development transitions took place at the company shareholders meeting leadership roles, assumed the position of director of busi- on March 23, 2017. ness development. With this change, Kitware consolidated its business development, proposal writing, and communica- At the meeting, Kitware named former Vice President of tions teams under Davis. The new organization will better Commercial Operations Lisa Avila, Ph.D., as president and align companywide strategic growth, proposal writing, and CEO. Avila co-founded Kitware in 1998. Throughout the last marketing goals. two decades, Avila has influenced every aspect of the busi- ness from technical development and project management Kitware reflected these and other transitions on its website. to business development and communications. For inquiries related to Kitware, its leadership, and its tech- nologies, please contact kitware(at)kitware(dot)com. “I am proud that our innovative open-source software has advanced scientific research, improved healthcare outcomes, KITWARE AND NEWMONT GUIDE MINING and enhanced our national security,” Avila said. “I look WITH VIRTUAL REALITY forward to maintaining our commitment to our customers, Through a combined effort, Kitware and Newmont Mining collaborators, communities, and team members while fos- Corporation (Newmont) visualized ongoing expansion of tering continued growth by expanding into new domains.” the Tanami mine in virtual reality (VR). For the work, Kitware added functionality to ParaView, its open-source software Former President and CEO Will Schroeder, Ph.D., another solution for high-performance computing (HPC). Kitware co-founder, began a new role as opportunity cata- lyst. In this role, Schroeder will work across the company on “We teamed up with Newmont geologists to not only model mentorship, technical leadership, business development, the existing underground development of the Tanami mine and customer engagement. He will also continue to provide but to map planned activities in accordance with results
10 from exploration drilling,” said Ken Martin, who led the load and save viewpoints have allowed Newmont geologists visualization effort at Kitware. “The virtual environment in to quickly navigate the virtual mine model. Members of ParaView has helped geologists to investigate whether their Newmont demonstrated such navigation at this year’s BMO expansion plans will allow miners to reach gold in the most Capital Markets Annual Global Metals & Mining Conference, effective and efficient manner.” where they led attendees through Tanami with VR.
The Northern Territory of Australia holds the Tanami mine. “It has proven to be a powerful communication tool to To date, the mine includes around 80 miles of tunnels and showcase our operations and projects to investors, analysts, reaches over 5,000 feet below the surface of the Earth. reporters, and Board members,” Godoy said. According to Newmont’s Group Executive for Resource The Visualization Toolkit (VTK), which provides the rendering Modeling, Marcelo Godoy, over the next two years, Newmont engine for ParaView, will gain the enhanced functionality plans to complete approximately 76 miles of core drilling, that Kitware worked with Newmont to produce in the 8.0 aiming to increase estimation confidence and expand the release cycle. gold reserves. KITWARE SHARES RESULTS OF DENTAL “The work we do in mining makes intense use of three- SHAPE ANALYSIS RESEARCH dimensional visualization of spatial data, so we have been Kitware shared preliminary results of ongoing research at monitoring advances in VR technology for a few years,” the 95th General Session & Exhibition of the International Godoy said. “The traditional mining software providers were Association for Dental Research (IADR). Kitware Technical too slow to include VR into their development roadmaps; Leader Beatriz Paniagua, Ph.D., presented the results in therefore, we decided to partner with Kitware to bring the “Continuous 4D Shape Analysis of Mandibular Changes” technology to a level where it can be effectively used for at an oral session on temporomandibular joint disorders resource modeling and mine planning.” (TMD). The session, TMD Imaging Advances in Craniofacial Biology and Orthodontics, took place Friday, March 24, 2017, in Moscone West in San Francisco, California.
The research that Paniagua presented aims to help dental scientists characterize the type of structural changes that result in the mandible from orthognathic surgery. This research showcases an example application of shape regres- sion that Kitware and team members plan to incorporate into an open-source software framework in a multi-year initiative. The initiative began in December 2016, when the National Institutes of Health awarded the team with a Research Project (R01) grant. Since this time, team members have worked to integrate existing shape analysis research code into the framework, which they named Slicer Shape AnaLysis Toolbox (SlicerSALT). Once completed, SlicerSALT will become available as a customized version of 3D Slicer. A screenshot from a video depicts Director of Business “SlicerSALT will allow researchers who may not have Development Brad Davis as he looks at the Tanami mine in virtual reality. Kitware published the video at https://vimeo. expertise in shape analysis to evaluate the geometrical prop- com/212599095. erties of different three-dimensional anatomical structures,” Paniagua said. “Our team’s results demonstrate the usability To assist Newmont geologists, a team from Kitware added of SlicerSALT to characterize postsurgical conditions such as functionality that grants users control over the exact scale of temporomandibular joint resorption.” the virtual world in ParaView. This functionality has encour- Paniagua presented during the 2:30 to 2:45 p.m. PST time aged Newmont geologists to easily transition from viewing slot. She also participated in the Using CBCT for Measuring the entire model of the mine to individually touring life-size Temporo-Mandibular Joint Changes and Tooth Movement tunnels and branches. “VR provides a unique way for us to workshop, which occurred Saturday, March 25, 2017, from interact with our drilling data, geological models, and mine 2 to 3:30 p.m. PST. plans,” Godoy said. “The ability to visualize huge amounts of data at real-world scale enhances our abilities to recognize Research reported in this publication was supported mineralization patterns and make better design decisions.” by the National Institute Of Biomedical Imaging And Bioengineering of the National Institutes of Health under The Kitware team worked with Newmont to create a cus- Award Number R01EB021391 and by the National Institute tomized version of ParaView that enables users to virtually Of Dental & Craniofacial Research of the National Institutes travel at scale-appropriate speeds. Further refinements to
11 of Health under Award Number R43DE024334. The content Lucas Gandel is solely the responsibility of the authors and does not neces- Gandel rejoined the team in Lyon, France, as an intern. sarily represent the official views of the National Institutes Xiaochuan (Matthew) Ma of Health. Ma joined Kitware as an R&D engineer in data and analytics KITWARE HIRES DOUBLE DIGITS IN 2017 in February 2017. So far this year, Kitware has recruited 20 new team members. Jean-Baptiste Vimort The company has posted information on open opportunities Vimort began an internship in medical computing. He at http://jobs.kitware.com/opportunities.html. attends CPE Lyon to pursue his master's degree.
Sam Gerber Neal Siekierski Gerber brought experience with fast algorithms and large Kitware added Siekierski to its Carrboro office. Siekierski's biological datasets to the medical computing team in prior work involved convolutional neural networks and Carrboro, North Carolina. retinal image segmentation.
Laura Pascal Mmanu Chaturvedi Pascal studies image processing and software development From Colorado State University, Chaturvedi came to the at École Supérieure de Chimie Physique Électronique de Lyon computer vision team as an intern. (CPE Lyon). She accepted a research and development (R&D) internship at Kitware. Rachel Clipp Clipp, whose many areas of expertise emphasize mechanical Bo Dong and biomedical engineering, started in Carrboro as a Staff Kitware welcomed Dong to the computer vision team as a R&D engineer. senior R&D engineer. Jeff Webb Wayne Durr Web commenced his role as an R&D engineer in March 2017. Durr took on the role of contracts administrator at the His work at Kitware relates to both medical computing and company headquarters. He holds an MBA. computer vision.
Brian Wylie Aaron Bray Wylie became a technical leader. His research interests The medical computing team grew with the addition of include information security. Bray, who went to Auburn University.
In addition to providing readers with updates on Kitware Contributors: Aashish Chaudhary, Brad Davis, Gabor product development and news pertinent to the open Fichtinger, Andras Lasso, David Manthey, and Csaba Pinter source community, "Kitware Source" delivers basic information on recent releases, upcoming changes, and Graphic Design: Steve Jordan technical articles related to Kitware open-source projects. Editor: Sandy McKenzie For an up-to-date list of Kitware projects and to learn about areas into which the company is expanding, please visit This work is licensed under an Attribution 4.0 https://www.kitware.com. International (CC BY 4.0) License.
A digital version of "Kitware Source" is available in a blog Kitware, ParaView, CMake, KiwiViewer, and VolView are format at https://blog.kitware.com/the-source-newsletter. registered trademarks of Kitware, Inc. All other trademarks Kitware would like to encourage members of its active are property of their respective owners. developer community to contribute to "Kitware Source." Contributions may include a technical article that describes an enhancement made to a Kitware open-source project or successes/lessons learned via developing a product built on one or more Kitware open-source projects. "Kitware Source" is published by Kitware, Inc., Clifton Park, New York.
12