Multimedia Information Extraction Roadmap

Gregory K. Myers, Gökhan Tür, L. Lynn Voss, Bob Bolles, Sachin S. Kajarekar, Elizabeth Shriberg, Dilek Hakkani-TürAAAI 2008 · OpenReview

Abstract

Multimedia Information Extraction Roadmap Greg Myers, Gokhan Tur, Lynn Voss, Bob Bolles, Sachin Kajarekar, Elizabeth Shriberg, Dilek Kakkani-Tur Multimedia Information Extraction Roadmap Critical Technical Challenges Information about people, their activities and communication is one of the most important types of content to extract from video data. This information is expressed in both the audio and visual domains. The critical technical challenges for extracting such content include: 1) Understanding interactions between people — their relationships, functional roles, hierarchies and dominance; and understanding their activities. 2) Broadening the robustness of multimodal information extraction techniques beyond narrowly constrained sets of actions, limited venues, and reliance on highly instrumented collection environments. 3) Obtaining sufficient amounts of annotated data for training models and classifiers. 4) Developing novel multimodal fusion techniques for semantically complex tasks involving human behavior. Submitted: Sep 12, 2008
BibTeX
@inproceedings{myers2008multimedia,
  author = {Gregory K. Myers and Gökhan Tür and L. Lynn Voss and Bob Bolles and Sachin S. Kajarekar and Elizabeth Shriberg and Dilek Hakkani-Tür},
  title = {{Multimedia Information Extraction Roadmap}},
  booktitle = {AAAI Conference on Artificial Intelligence},
  year = {2008},
  url = {https://openreview.net/forum?id=V56F-kJfUTU},
}

Loading related content...

Figure