Der deutsche Education Blog

January, 2014

Microsoft Research Outreach Blog

The Microsoft Research Outreach blog shares stories of collaborations with computer scientists at academic and scientific institutions to advance technical innovations in computing, as well as related events, scholarships, and fellowships.

January, 2014

  • Microsoft Research Outreach Blog

    Stroke recovery gets a boost from Kinect


    The aftermath of a stroke can be overwhelming for any patient—from the physical and emotional toll to the cost of stroke-related treatment. Recent evidence1 points to the advantages of task-specific training as effective rehabilitation, but in practice, this requires simple, repetitive movements, which may bore patients, lowering their motivation to continue the training. However, new hope for stroke patients has arrived in the form of Stroke Recovery with Kinect, a research project to build a cost-effective, interactive, home-rehabilitation system for motor recovery after a stroke—based on Microsoft Kinect technology.

    Kinect offers a new approach to physical therapy

     Stroke Recovery with Kinect is a collaborative project between Microsoft Research Asia and Seoul National University—with funding from the Korean Government Collaboration Program—that provides a virtual reality system to help stroke survivors improve their upper-limb motor functioning in the comfort of their own home. “Most people who suffer a stroke experience paralysis in their arms and legs,” states Professor Nam-Jong Paik of Seoul National University, who is principal investigator of the project. “They can do the therapy at home by using Microsoft’s Kinect—without coming to the hospital—and we can measure their recovery level. Since it’s like a game, patients also have fun while rehabilitating at home.”

    The prototype Stroke Recovery with Kinect system was built by using the Microsoft Kinect for Windows software development kit (SDK). It uses the Kinect sensor’s three-dimensional camera to capture the movements of 48 skeletal points on the patient while he or she performs the therapy. Stroke Recovery with Kinect interprets the movement data, enabling the system to measure and evaluate the patient’s movements and assess their rehabilitation progress. The system uses the patient's scores from previous sessions to adjust the level of difficulty for subsequent therapy sessions.

    One of the three programs in the Stroke Recovery with Kinect system is the classic box-and-block test (BBT). This program application evaluates patients’ coordination, gross manual dexterity, and motor skills as they (virtually) attempt to pick up blocks one-by-one and put them into a box in a set amount of time. Similar to a computer game, Stroke Recovery with Kinect displays patients’ scores as soon as they finish a session, providing immediate reinforcement when scores improve from session to session.

    The box-and-block test in Stroke Recovery with Kinect evaluates a patient’s coordination, manual dexterity, and motor skills.
    The box-and-block test in Stroke Recovery with Kinect evaluates a patient’s coordination, manual dexterity, and motor skills.

    Another program in Stroke Recovery with Kinect challenges the patient to assume a target pose displayed on the computer monitor and then duplicate the target’s position as it moves. The patient then receives what is known as a Fugl-Meyer Assessment (FMA) score, based on his or her success. Because Stroke Recovery with Kinect enables patients to face these challenges within the privacy of their own homes, they may be more relaxed and likely to persevere.

    The third program in Stroke Recovery with Kinect is an outer-space game that enables patients to exercise their reflex and reaction abilities as they guide a spaceship through space while attempting to avoid oncoming asteroids. Stroke Recovery with Kinect tracks the stroke patient’s hand trajectory—relative to and in conjunction with the movement of the elbow and/or shoulder. The stroke patient experiences a fun and enjoyable therapy session that a traditional rehabilitative setting usually cannot provide.

     Long-term plans for Stroke Recovery with Kinect include integrating social networking into the system so that stroke patients can connect with one another and participate jointly in the rehabilitative programs, building a sense of camaraderie that could offer emotional and psychological support and motivation. Within the community, patients will have the opportunity to communicate about their condition and receive encouragement as they advance toward recovery. Future updates will make it possible for doctors to monitor the patient’s rehabilitation from the hospital or their office, and to communicate with the patient regarding their treatment and progress. Additionally, as the system becomes more widely used, we anticipate incorporating machine learning into the system.
    Finally, this home-based rehabilitation system also has potential cost benefits. The expense of ongoing stroke-related office visits for rehabilitation burdens healthcare systems and patients worldwide.

    I look forward to future collaborative efforts between Microsoft Research Asia and Seoul National University on the Stroke Recovery with Kinect project. We expect Stroke Recovery with Kinect to pave the way for stroke patients to save both time and money through a convenient, effective, and enjoyable rehabilitation program.

    —Miran Lee, Senior Manager, Microsoft Research Connections

    Learn more

    1Kleim JA, Jones TA: "Principles of experience-dependent neural plasticity: implication for rehabilitation after brain damage," J Speech Lang Hear Res 2008, 51:225-39.

    Return to blog


  • Microsoft Research Outreach Blog

    Latest recipients of Windows Azure for Research Awards announced


    Microsoft Research’s Windows Azure for Research program, which features a continuing series of Windows Azure cloud training events and a program of Windows Azure research grants, has been going strong since its launch in September 2013. As the December 15, 2013, deadline for the second round of grant proposals approached, we braced ourselves for a barrage of creative ideas. We weren’t disappointed, receiving proposals from every continent (well, except Antarctica). The response was particularly strong from such countries as Brazil and China, where our recent training events gave researchers an excellent, hands-on view of the capabilities of Windows Azure.

    Forty-five proposals selected from researchers around the world

    Several strong research themes that had emerged in the first round of proposals continued in the second round. Specifically, the life sciences and the emerging field of urban science were abundantly represented. Both themes can be thought of as big data topics, but they are really part of what we call the fourth paradigm of science, which is about discovering new scientific principles through deep analysis of massive amounts of data.

    Urban science, which can be described as an interdisciplinary mash-up of computer science and social science, is becoming an important tool for city planners. By using the real-time data that a typical modern city generates, they can gain a better understanding how to improve life for the city’s inhabitants. The cloud is ideally suited to collecting, filtering, analyzing, and sharing these data.

    A set of related topics that came on strong in the second-round proposals involved environmental science, ecology, and geosciences. Again, the common theme is using Windows Azure on the Microsoft cloud for data collection, analysis, and dissemination. In addition to such fourth-paradigm ideas, we received a large number of excellent computer science proposals that rely on the scale of the cloud to experiment with new algorithms and database topics.  

    Selecting the winning proposals was extremely difficult, as we can fund only a fraction of the submissions. Nonetheless, we persevered and winnowed the proposals down to the grant recipients listed, by lead author and project title, at the bottom of this blog. The order might appear random, but trust me, there’s a logic to it (hint: take a look at the alphabetical order of the country names). You can review abstracts for these proposals at Windows Azure for Research.

    As a reminder, the next deadline for proposals is February 15, 2014. We encourage potential applicants to attend one of our training events or, if that’s not possible, to study the training material we’ve posted online. You can find a schedule of upcoming training events and the aforementioned training materials at Cloud Research Projects.

    Dennis Gannon, Director of Cloud Research Strategy, Microsoft Research Connections

    Learn more

    Second-round Windows Azure for Research Award recipients:

    • Jian Zhang, University of Technology, Sydney, Australia
      Friends Recommendation Based on Graph Correlation
    • Yuedong Yang, Griffith University, Australia
      Cloud-based Platform for Genome-scale Prediction of Protein Functional Complex Structures at Experimental Quality    
    • Altigran Soares da Silva, UFAM, Brazil
      Keyword-based System for Relational Database    
    • Carmem Satie Hara, Federal University of Parana, Brazil
      RING Project    
    • Fernando da Fonseca de Souza, Universidade Federal de Pernambuco, Brazil
      Cloud Databases: A model to guarantee data consistency    
    • Luiz André Portes Paes Leme, Universidade Federal Fluminense, Brazil
      Assessing Recommendation Approaches for Dataset Interlinking    
    • Marcelo Valadares Galdos, Brazilian Bioethanol Science and Technology Laboratory (CTBE) / Brazilian Center of Research in Energy and Materials (CNPEM), Brazil   
      Using Azure to run an integration of process-based environmental models and geographic information systems    
    • Marta Mattoso, Federal University of Rio de Janeiro, Brazil
      User-Steering Phylogenetic Workflows in the Cloud
    • Milton Cezar Ribeiro, Sao Paulo State University, Brazil
      Integrating phonological, landscape, fauna movement and remote sensing massive data and processing throughout e-Science and Cloud computing
    • Rafael Duarte Coelho dos Santos, INPE - Brazilian National Institute for Space Research, Brazil
      Prototype Deployment of a Data Server for the Brazilian Weather and Climate Virtual Observatory    
    • Ricardo da Silva Torres, Institute of computing, University of Computing, Brazil
      Big Image Data Management on the Cloud for e-Science Applications    
    • Guangjun Zhang, Peking University, China
      Machine learning – parameter estimation for groundwater flow and transport models based on Windows Azure Cloud    
    • Huayi Wu, Wuhan University, China
      Collaborative Geoprocessing on Windows Azure
    • Jitao Sang, Chinese Academy of Sciences, China
      Cyber-Physical Footprint Association: Cloud Storage and Computing    
    • Junjie Wu, Beihang University, China
      A System for Heterogeneous Social Media Big Data Analytics in Azure Cloud    
    • Lei Zou, Peking University, China
      Graph Data Management in Urban Computing
    • Xinbo Gao, School of Electronic Engineering, China
      Videos analysis and recommendation for online learning    
    • Yan Xu, Beihang University, China
      Large-scale histopathology image analysis for colon cancer in Azure
    • Yuan Juli, Zou Hengming, Shanghai Jiao Tong University, China
      A Distributed Algorithm for File Distribution and Replication on Cloud Platform
    • Andres M. Pinzon, Center for Bioinformatics and Computational Biology of Colombia, Colombia
      A cloud-based system for the integration of molecular data and biodiversity information for Colombian species    
    • Frederic Magoules, Ecole Centrale Paris, France
      Advanced Linear Algebra Libraries for the Cloud    
    • Jean-Charles Régin, University Nice-Sophia Antipolis, France
      Using Windows Azure for High Performance Computing    
    • Liliana Pasquale, University of Limerick, Ireland
      Minority Report: Using the Cloud to Enable Proactive Digital Forensic Investigations
    • Tony Tung, Kyoto University, Japan
      Life Maps    
    • Hwasoo Yeo, Korea Advanced Institute of Science and Technology, Korea
      Cloud Sensing based Urban Travel Time Prediction with Online Traffic Simulator
    • Hyunju Lee, Gwangju Institute of Science and Technology, Korea
      Text mining for identifying disease-gene-biological relationships    
    • Joon Heo, Yonsei University, Korea
      Does ‘Gangnam Style’ really exist? - Answers from data science perspective    
    • Muhammad Bilal Amin, Kyung Hee University, Korea
      Enabling Data Parallelism for large-scale Biomedical Ontology Matching over Multicore Cloud Instances
    • Grigor Aslanyan, University of Auckland, New Zealand
      Studying Very Early Universe Physics with Cosmic Microwave Background Anomalies
    • Marek Stanislaw Wiewiorka, Warsaw University of Technology, Poland
      Towards an interactive secondary analysis of RNA sequencing data service in Widows Azure cloud with Apache Spark framework
    • Heiko Schuldt, University of Basel, Switzerland
      ADAM+ - A Large-Scale Distributed Image and Video Retrieval System
    • Blesson Varghese, University of St Andrews, United Kingdom
      Real-time Catastrophe Risk Management on Windows Azure    
    • Julio Hernandez-Castro, David Barnes, University of Kent, United Kingdom
      ChessWitan: Mining chess data to distinguish human from computer play
    • Nadarajen Veerapen, University of Stirling, United Kingdom
      Automated Bug Fixing
    • Vassilis Glenis, Newcastle University, United Kingdom
      Modelling Flood Risk in Urban Areas    
    • A. Lucas Stephane, Florida Institute of Technology, United States
      Life-Critical Interactive Glass Wall Integration    
    • Alexander Vyushkov, University of Notre Dame, United States
      Modeling Malaria Transmission on Windows Azure    
    • David Hazel, University of Washington, United States
      AMADEUS - Azure Marketplace of Applications for Diverse Environmental Use as a Service
    • Dhruv Batra, Virginia Tech, United States
      CloudCV: Large-Scale Distributed Computer Vision as a Cloud Service    
    • Hanspeter Pfister, Harvard University, United States
    • Kelly Smith, University Corporation for Atmospheric Research (UCAR), United States
      The Unidata Integrated Data Viewer (IDV) as a Cloud Service
    • Richard Dana Loft, National Center for Atmospheric Research, United States
      AzurePlanet: A cloud-based system providing access to weather and climate information
    • Susan Borda, California Digital Library, United States   
    • Tanya Berger-Wolf, University of Illinois at Chicago, United States
      Computational Behavioral Ecology on the Cloud    
    • Yuejie Chi, The Ohio State University, United States
      Online Distributed Inference of Large-Scale Data Streams in the Cloud   

    Back to blog

Page 3 of 3 (8 items) 123