Exploiting Visual Cues in Non-Scripted Lecture Videos for Multi-modal Action Recognition.
Ali Shariq ImranAlejandro MorenoFaouzi Alaya CheikhPublished in: SITIS (2012)
Keyphrases
- multi modal
- action recognition
- visual cues
- lecture videos
- low level
- human actions
- visual information
- activity recognition
- mid level
- bag of words
- computer vision
- learning materials
- semantic concepts
- key frames
- multiple modalities
- image annotation
- domain knowledge
- high dimensional
- similarity measure
- high level
- user experience
- learning objects
- hybrid learning
- image processing