Upload
claude-bond
View
212
Download
0
Embed Size (px)
Citation preview
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
Integrated Media Systems Center
Ulrich Neumann, DirectorCharles Lee Powell Professor of Engineering
Associate Professor, Computer Science
Isaac Maya, Ph.D., P.E. Director, Industry and Technology Transfer Programs
IMSC’s Immersive Technologies
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
NSF
Integrated Media Systems Center
a partnership among:National Science FoundationUniversity of Southern California
School of Engineering Annenberg Center for
CommunicationState of CaliforniaThe City of Los AngelesIndustry Partners: Computer Hardware and Software Entertainment Broadcasting Telecommunications Publishing Aerospace
Other Government Agencies:DARPA, NASA, JPL, ONR, US Army
$10 M operating budget
Downtown LA
IMSC Headquarters
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
Faculty and Academia
A top-notch array of investigators – 14 of whom (~ 50%) were sought specifically because of IMSC
29 investigators working with 87 Phd, 26 MS, and 33 UG students
Investigators come primarily from EE and CS – others from Psychology, Industrial and System Eng., School of Cinema/Television, Annenberg School for Comm., School of Gerontology, Biomedical Engineering, and the Information Sciences Institute
Two IMSC investigators have PYI awards Seven IMSC faculty have CAREER awards (3 CAREER awards this year) Tomlinson Holman won a 2001 Technology Achievement Award from the Academy of
Motion Pictures Robert Sholtz won a MILCOM Achievement Award for UWB research (only 4th in 20 yrs)
IMSC faculty published 53 peer-reviewed journal articles and 205 peer-reviewed conference papers
articles appearing in print over a 12 month period in this past year (2001-2002)
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
Education
Programs created by IMSC
171 students graduated with IMSC providing funding, classes, and research aspects of their education experience
81 with PHD, 77 with MS, and 13 with BS
IMSC created 3 MS programs enrolling 358 students unique among U.S. universities
IMSC created 2 UG minor programs enrolling 207 students
IMSC gave research fellowships to 32 UG students
Created more than 20 courses for programs – e.g.: Human Factors in Integrated Media Systems Integrated Media Systems and Architectures School of Engineering/Fine Arts (joint project course)
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
Industry Program IMSC offers cost-effective research partnership at the 5-10 yr horizon
Pursue science and discovery of methods and algorithms – IP licenses Develop technology-impact and proof-of-concept before “start up” Prototype feasibility applications and testbeds Provide partners a view to potential opportunities and threats
Strong support through six years 28 members this past year (2001-2002)
Recognized 5 members for consistent partnership over 6 yr life of center HP, NCR, FXPAL, TRW, Lockheed Martin
Strong growth in international corporations and government programs Taiwan (III, ITRI) Korea (LG Electronics) Japan (FXPAL, Hitachi, NTT DoCoMo, Toshiba) Netherlands (Philips Research) Italy (ST Microelectronics)
Patents and tech transfer 46 patents filed over the life of center 24 patents licensed and 39 software licenses
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
IMSC Research Members
IIIITRI
Taiwan
Microsoft WA
See California Map
Collaborations with Government Agencies and Foundations:
US ArmyDARPANASA Toyota Foundation
NCR OH
Eastman Kodak NY
DRS TechnologiesPrentice Hall PTR
NJ
NTT DoCoMo Japan
NavTech IL
Intel OR
Time Domain AL
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
Silicon Valley
Hollywood
Los Angeles
Silicon Valley
•ConceptLabs•F-X Palo Alto Laboratory•Geometrix•Hewlett-Packard•InterVideo•Lockheed Martin•Sun Microsystems
USCIMSC
Los Angeles/San Diego
• AboveNet Communications• Boeing• HRL• JPL• Los Angeles Times• NMUA• Panoram Technologies• TMH Corporation• TRW
IMSC Research Members
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
IMSC Research Program -- ImmersipresenceMotivated by Moore’s law: exponential performance growth...
enables advances towards the future of the internet, grid computing, nano-technology, ...
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
People are often the ultimate purpose behind silicon systems
In pursuit of an experience, or the performance of a task, or understanding of information
Integrated Media System (IMS) mediate the gap between silicon and humans
An extension of the human senses, actuators, or cognitive process
Our goal is to advance the science and engineering of “well designed” Integrated Media Systems that exploit human information processing capabilities - providing effectiveness that we associate with “immersipresense”.
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
To better understand immersipresence and help set
specific goals, we develop visions of how IMS
technologies will affect people and their activities
We examine four domains for IMS impacts in the future
These four vision scenarios drive the center’s program of six research and engineering areas
UWB console captures 3D model and haptics scan
PDA provides spatial array of participants and sonic
translation of haptic trace
Immersive workstation coordinates video, avatars, graphics, audio, and haptics to facilitate problem resolution
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
Immersive Audio
Facial Gesture Analysis and Animation
Robust Vision Analysis
Immersi-data Content Extraction & Analysis
Theory of Perceptual and Cognitive Pleasure
User State Sensing and Perceptual User Interfaces
Virtual Environments for Performance Testing and Training
Multimedia Networks, Transmission, and Comm-unication
Ultra Wideband Wireless
Remote Media Immersion
Software Architecture for Immersi-presence
HAPTICS – devices and translations
Immersive on-demand streaming of audio/video experiences - fusion of cinema and internet
Virtual/video peer-to-peer telepresence experiences for
multiple participants
Full body immersion role-playing game experiences with realistic or synthetic characters
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
Immersive Audio
Speech Recognition and Synthesis
Facial Gesture Analysis and Animation
Robust Vision Analysis
Digital Geometry Processing
Immersive Media Real-Time Storage & Retrieval
Emotional Expressions
User-State Sensing and Perceptual User Interfaces
Multimedia Networks, Transmission, and Comm-unication
Ultra Wideband Wireless
Remote Media Immersion
Software Architecture for Immersi-presence
Remote Media Immersion - fusion of internet and cinema
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
Theory of Perceptual and Cognitive Pleasure
Emotional Expressions
User State Sensing and Perceptual User Interfaces
Virtual Environments for Performance Testing and Training
Information Integration
Immersi-data Content Extraction & Analysis
Customized Querying & Rendering
Immersive Media Real-Time Storage & Retrieval
Remote Media Immersion
Software Architecture for Immersi-presence
RMI - fusion of internet and cinema
BioSIGHT - authoring and assessment
I-NEWS - story models and user preferences
HAPTICS - devices and translations
Immersive audio
Speech Recognition and Synthesis
Facial Gesture Analysis and Animation
Robust Vision Analysis
Digital Geometry Processing
Multimedia Networks, Transmission, and Comm-unication
Ultra Wideband
Wireless
Compression
IMSC research program
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
Research Highlights
immersive audio multichannel and HRTF approaches - holistic DSP approach
streaming servers distributed and scalable architecture
UWB wireless leadership in technology, FCC regulation, and commercialization
perception & cognitive modeling theoretical foundations for understanding benefits of media immersion
computer vision computational framework for grouping based on tensor voting
graphics & animation
3D DSP mesh processing, facial expression analysis and animation, hair modeling
virtual reality -- chair of VR2003 in LA applications to psychology (ADD diagnosis), haptics applications, and user studies
IMSC has produced ground breaking results and fundamental research in:
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
IMSC Immersipresence Vision: Applications in Entertainment
Animation File Goes Here:PC – panoramic_animation.aviMAC - panoramic_animation.mov
360º Camera
USC Homecoming Game at the Los Angeles Coliseum
Panoramic Visual Immersion: Malibu Hills
QuickTime™ and aCinepak decompressor
are needed to see this picture.
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
IMSC Immersipresence Vision: Medical Applications
Animation File Goes Here:
PC – 200_0003.AVI
Evaluation of ADHD in Children
QuickTime™ and aMicrosoft Video 1 decompressorare needed to see this picture.
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
Video FilesMac: face_1.movPC: avatars_doug.avi
Animation and Modelingof People for Immersipresence
3D Visual Environment
Avatar simulates live video
Avatar: Digital Cloning
QuickTime™ and aCinepak decompressor
are needed to see this picture.
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
Animation and Modeling:Expression Cloning
Real Expression can be “cloned” to any model
QuickTime™ and aNone decompressor
are needed to see this picture.
QuickTime™ and aNone decompressor
are needed to see this picture.
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
Video FilesMac: face_1.movPC: avatars_doug.avi
Animation and Modeling:Expression Analysis
Real Expression can be “cloned” and analyzed
QuickTime™ and aNone decompressor
are needed to see this picture.
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
Digital Geometry Processing: 3D Modeling and Elasticity
Animation File Goes Here:PC – mat_1.aviMAC - mat_1.mov
Animation File Goes Here:PC – mat_2.aviMAC - mat_2.mov
Animation File Goes Here:PC – mat_3.aviMAC - mat_3.mov
Next wave of visual data: Better understanding and manipulation of geometric data
QuickTime™ and aCinepak decompressor
are needed to see this picture.
QuickTime™ and aCinepak decompressor
are needed to see this picture.
QuickTime™ and aCinepak decompressor
are needed to see this picture.
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
Mixed Reality and Visualization
Augmented Reality for Space Flight (NASA funded)
Develop AR authoring tools for video-based training
Anthony Majoros, Human Factors group, The Boeing Corp.
4D Battlefield Visualization (MURI funded)
Develop fusion of video, images, and 3D models Avideh Zakhor (Berkeley), Suresh Lodha (UC
Santa Cruz), Bill Ribarsky (Georgia Tech), Pramod Varshney (Syracuse)
Wide Area AR Tracking (ONR funded)
Novel sensors and fusion for tracking position/orientation
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
Non-negative Negative
(From the same person and dialog)
Recognizing Emotions in Speech
Natural Human-computer interaction: need for machines that can recognize “what, who and how” Call center applications, Teleconferencing, Tutoring by
machines, Entertainment Algorithms (ASRU’01, ICME’02, ICSLP’02)
Support Vector Machine, Linear discriminant and k-NN classifier for acoustic information
Novel information-theoretic measure of emotionalsalience in language (word sequences)
Acoustic and linguistic information fusion Performance evaluated on real call center data
rather than feigned data from actors
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
Video FilesMac: audio_10.2_channel.movPC: audio_10.2_channel.avi
Immersive Audio 10.2
3D Virtual audio field at the location of the ears at all times
Immersive Audio Environment
QuickTime™ and aCinepak decompressor
are needed to see this picture.
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
Haptics Research
Force control for haptics rendering Collision detection Compression of haptic data Haptic playback and analysis Application Research Project: Haptic Museum
CyberGrasp PHANToM (3D Mouse)
Movie File:PC: haptic_2.aviMAC: haptic_2.mov
Movie File:PC: haptic_1.aviMAC: haptic_1.mov
QuickTime™ and aCinepak decompressor
are needed to see this picture.
QuickTime™ and aCinepak decompressor
are needed to see this picture.
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
Haptics Experiment
Virtual “touching” through the internet Virtual “handshaking” Distributed Haptic World
INTERNET
Mutual Touch
CyberGrasp PHANToM
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
Media CommunicationsHigh Bandwidth Networks: synchronization, latency, bandwidthHigh Bandwidth Networks: synchronization, latency, bandwidth
Wireless SystemsWireless Systems
Optical CommunicationOptical Communication
CompressionCompression
Photonic NetworksPhotonic Networks
QoS Management of IP NetworksQoS Management of IP Networks
Encryption and Network SecurityEncryption and Network Security
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
Managing ImmersidataHaptic Application
Control
Immersive Audio Text3D Model
Avatar
Video
Immersidata
Real-time acquisition and playback Efficient query and analysis Integration of disparate information sources
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
With the aim to With the aim to improve:improve:
Effectiveness Efficiency Enjoyment Safety
Human Factors Human Computer Interaction General Experimental Psychology Usability Engineering Communication User-Centered Design Cognitive Science Ethnography
“User Centered Sciences” Research
Integrates Multiple Methodologies:
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
Systems Engineering
RMI testbed for streaming media immersionShowcase for IMSC research and integrationMay 2002 demonstration (NY Times)
Flow Scheduling Framework of the Media Immersion Environment - open source release
BioSIGHT authoring tools and web-based educationTesting and assessment in K-12 classrooms
Similarly, unique systems engineering and integration are hallmarks of IMSC
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
RMI A fusion of internet browsing with a theater-like
immersive experience HD Video at up to 45 Mbits/sec 10.2 ch Immersive audio (16 Mbits/sec)
Steaming over the Internet on-demand to a mouse click (20 min prog)
RMI demonstrates IMSC innovations in several critical technical areas:
Immersed in a college football game
Doctors assisting in a remote procedure
Business people negotiating like they
are in the same room
Students visiting an aquarium a thousand
miles away
Streaming media
servers
Immersive audio
capture and
rendering
Network protocols for error
correction Synchronization
Applications
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
RMI Challenges Immersive, high-quality video and
audio acquisition and rendering High Definition video 1080i and
720p (~40 Mb/s) 10.2 channels of uncompressed
audio (12 Mb/s)
Storage and transmission of media streams across networks Streaming media
servers/clients Error correction via selective
retransmission Synchronization between
streams A/V, A/A, V/V
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
RMI Architecture
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
Storage, Streaming & Rendering Distributed
Server Storage Scheduling Scalability
Clients Multi-stream
Synchronized playback
Transmission Robust VBR
Flow control
Requirements:
A streaming platform that can scale and handle synchronized, high-bandwidth streams
Focus:
End-to-end streaming architecture
[ACM Computer’02]
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
Scalability: Multi-Node, Multi-Disk Data and control network traffic can be routed with
different logical topologies Yima-1: single data path
(high inter-node traffic) Yima-2: multiple data paths
(low inter-node traffic)
Yima-2Yima-1
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
Robust Stream Delivery
Variable bit rate (VBR) media encoders allocate more bits to complex scenes and less bits to simple ones
Smoothing of VBR media traffic has the following quality benefits: Better resource utilization (less bursty) More streams with the same network capacity
Multi-Threshold Flow Control(MTFC) algorithm objectives: Online operation Content independence Minimizing feedback
control signaling Rate smoothing
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
Premier Demonstration May 9, 2002
Cross-country streaming from ISI-East in Virginia
NY Times coverage in “Circuits” section
NBC-TV and KTLA http://imsc.usc.edu/rmi/ http://www.east.isi.edu/NGI-
S/
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
Future Impact of RMI High-quality audio/video experience streamed on-
demand over the internet (live or recorded) Everyone is a content producer, distributor, theater
Currently only broadcast, cable, theater, or satellite operators can distribute content…
high entry cost and regulation barriers RMI technology removes distribution barriers
CEO NVidia: “the end of TV as we know it” LIVE concerts, sporting events, alternative news
coverage, … RECORDED movie archives, news archives,
education programs, … “More than TV” experience – powerful human motivation
Playback quality and options are flexible – retain all the capabilities of computing on the internet
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
Distributed Immersive Performance Next generation of Remote Media Immersion (RMI) Vision – interactive peer-to-peer streaming
Create seamless immersive environment for distributed high-quality audio/video interaction
Impact – widespread telepresence Distributed performances, meetings, social
gatherings, education, Technical barriers
Precise timing and synchronization of streams Low latency, error correction – high quality-of-service
network protocols
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
N a t i o n a l S c i e n c e F o u n d a t i o n E n g i n e e r i n g R e s e a r c h C e n t e r
Summary IMSC’s vision of Immersipresence and Integrated Media
Systems aims for fundamental breakthroughs in technology, industry partnerships and education
IMSC is carrying out Highly cross-disciplinary cutting edge research Aggressive industry collaboration and technology
transfer Unique educational experience and teamwork
skills for students
IMSC is eager to expand its Collaborations
Acknowledgments
National Science Foundation Engineering Research Center program
http://www.eng.nsf.gov/eec/erc.htm
IMSC Industry partners.....
More information about IMSC
http/:imsc.usc.edu
IMSC PARTNERS Airborne 1 Corporation Compression Science CorporationConceptLabsEastman KodakF-X Palo Alto LaboratoryGeometrixHewlett-PackardHRLIBMInformixInstitute for Information IndustryIntelInterVideoJPLLittonLockheed MartinLos Angeles TimesLucent Technologies/Bell LabsNCRPanoram TechnologiesPhilips Multimedia CenterPrentice Hall PTRRaytheonSTMicroelectronicsSun MicrosystemsTMH CorporationTRW