Tao Huang Shrideep Pallickara Geoffrey Fox Community Grids Lab Indiana University Bloomington taohuang spallick gcfindianaedu Outline Analysis of existing Collaboration and Annotation Systems ID: 209832
Download Presentation The PPT/PDF document "A Distributed Framework for Collaborativ..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
A Distributed Framework for Collaborative Annotation of Streams
Tao Huang, Shrideep Pallickara, Geoffrey FoxCommunity Grids LabIndiana University, Bloomington {taohuang, spallick, gcf}@indiana.eduSlide2
Outline
Analysis of existing Collaboration and Annotation SystemsSystem ObjectivesSystem Design and ImplementationExperiment ResultsFuture Work and ConclusionSlide3
Collaboration Systems
Multimedia based collaborationIndustryAcademic Digital document based collaborationSlide4
Annotation Systems
Microsoft: MRASAsynchronous Q/A on lecture videosIBM : VideoAnnExMpeg-7 based annotationSupport distributed annotation University of Queensland : Vannotea Similar to VideoAnnEx with more flexibility on annotation metadataIndiana University CGL: eSports
Extend annotation forms from texts to images and video clipsAimed at distance coachingSlide5
Limitations of existing systems
MRAS and VideoAnnExAsynchronous communication and searching in annotation Vannotea and eSports Annotation on real time video streams in distributed environments None of them supports annotation on other types of data streams Slide6
Objectives
Support annotation on live or prerecorded data streams regardless their typesRelatively tolerant to possible failuresSupport saving, sharing and searching annotations among distributed users Easy to extendSlide7
System Architecture
Distributed annotation clients over the NaradaBrokering Network.Session management with two recovery strategiesDistributed stream archiving and replayingEasy to add support to handheld and streaming clients Slide8
Decoupled Functions in Layers
Presentation layerGUI, Stream Source/Renderer ManagersThe Logic layer Stream manager manages all active senders and receiversAnnotation manager associates and synchronizes content streams with annotation streams.The Transmission Creating and managing actual data transmissionSlide9
Stream Processing Interfaces
Define the generic behaviors of real time data streamsStart, Stop, Pause and so onExisting Stream Sources/RenderersAudio/VideoScreen captureFile captureEasy to write own source/renderer pairs Slide10
Annotation management
Annotation and Stream data are stored in remote storageAnnotation metadata is stored in plain XML filesOnly Time an Layout information are stored currentlyLocal Annotation metadata will be synchronized with remote storage before being used to replaySlide11
Client Interface
Stream Progress Panel
Stream List
Panel
Center Annotation
Panel
Client List Slide12
Preliminary Test 1Slide13
Preliminary Test 2Slide14
Conclusion
Woks well with data streams of small payloads(<2KB) under high stress circumstancesDelays are under a reasonable level, which will not cause any synchronization problemsSlide15
Summary
Support collaborative annotation on multimedia streams in Sync/Async wayStream Processing Interfaces make it easy to extend current systemSlide16
Future Work
Add more types of stream sources/renderersResearch on security issues in annotation distributionStandardize annotation metadata format into Mpeg-7 A user friendly Web 2.0 portal to access/manage annotation and stream dataSlide17
Questions?
Thanks