Multimodal Interaction Multimodal Interaction and TV Wei Yun Yau - - PowerPoint PPT Presentation

multimodal interaction multimodal interaction and tv
SMART_READER_LITE
LIVE PREVIEW

Multimodal Interaction Multimodal Interaction and TV Wei Yun Yau - - PowerPoint PPT Presentation

Multimodal Interaction Multimodal Interaction and TV Wei Yun Yau Programme Manager, Personalised Interactive Tele Services Institute for Infocomm Research, A*STAR Institute for Infocomm Research, A STAR Sheau Ng Vi Vice President &


slide-1
SLIDE 1

Multimodal Interaction Multimodal Interaction and TV

Wei‐Yun Yau

Programme Manager, Personalised Interactive Tele‐Services Institute for Infocomm Research, A*STAR Institute for Infocomm Research, A STAR

Sheau Ng

Vi P id t & H d f R&D Vice President & Head of R&D, NBC Universal

W3C Workshop on Rich Multimedia App Development

slide-2
SLIDE 2

Singapore

Institute for Infocomm Research (I2R)

2

slide-3
SLIDE 3

Who We Are

A f S i T h l & Agency for Science, Technology & Research (A*STAR) is under the Ministry of Trade & Industry, Singapore.

Vision Mission

Fostering world class

g p

Powering a vibrant and strong infocomm ecosystem for Fostering world-class infocomm and media research and developing a deep talent pool of i f f i l y Singapore. infocomm professionals to power a vibrant knowledge- based Singapore.

Singapore’s Largest ICT Research Institute g p

g

600

Staff today

681

Industry Collaborations 1220 Technology Disclosures 5430 Research Publications

Institute for Infocomm Research (I2R)

3

Institute for Infocomm Research (I2R)

3

slide-4
SLIDE 4

INTERNET

T t i l Terrestrial Satellite Cable DVD Player / Blue-Ray Player

Institute for Infocomm Research (I2R)

4

slide-5
SLIDE 5

Enjoying Related Content Seamlessly

Youtube News Local News station News

AlJazeera, AFP, Euronews

station (CNA)

Social Media

Institute for Infocomm Research (I2R)

5

slide-6
SLIDE 6

Challenges

  • Automated Content Discovery

– Where are the relevant multimedia content? – How relevance is determined?

  • User Preference Modeling

What the user likes? – What the user likes? – How to track content consumed from different sources on different devices?

  • Personalised Recommendation

– What the user would be interested now? – What are the external events happening that the user could be interested in?

  • Attractive Presentation

Attractive Presentation

– What to show, how to show, and when? – Easy interaction & navigation to create better engagement?

Institute for Infocomm Research (I2R)

6

Institute for Infocomm Research (I2R)

6

slide-7
SLIDE 7

News Video Analytics

External sources Story segmentation Named entities External sources Keyword Search / Timeline Speech‐to‐text segmentation Metadata enrichment Timeline News Video Tapestry Video‐OCR

Keywords & Summarization

Breaking News Alert Scene detection

… …

Link analytics

Linked data

Low-level Analytics

detection

High-level Analytics

Link analytics Local news channels

Applications

Linked data

y

channels Story-K Story-K+1 Story-K+2

Institute for Infocomm Research (I2R)

7

Institute for Infocomm Research (I2R)

7

slide-8
SLIDE 8

Demographics Social Interactions

Multiple Modality

Viewing Device

Demographics

Location Location

Mapping Contents to Viewers

Structured Semi‐structured / Unstructured

Institute for Infocomm Research (I2R)

8

Institute for Infocomm Research (I2R)

8

TV Content EPG ‐ Metadata Web Content Web Resources

slide-9
SLIDE 9

Personalised Recommendation

Questionnaire External Events Remote Usage Logs Remote Usage Logs

Social Groups

Feedback to Personalized EPG

Institute for Infocomm Research (I2R)

9

Institute for Infocomm Research (I2R)

9

Server: Contextual Input On Device: Daily Usage Input

slide-10
SLIDE 10

Thank you

Contact: wyyau@i2r a star edu sg Contact: wyyau@i2r.a-star.edu.sg

For additional information, please visit our website, htt // i2 t d / http://www.i2r.a-star.edu.sg/

Institute for Infocomm Research (I2R)

10