ACL RD-TEC 1.0 Summarization of N04-4023
Paper Title:
FEATURE SELECTION FOR TRAINABLE MULTILINGUAL BROADCAST NEWS SEGMENTATION
FEATURE SELECTION FOR TRAINABLE MULTILINGUAL BROADCAST NEWS SEGMENTATION
Authors: David D. Palmer and Marc Reichman and Elyes Yaich
Primarily assigned technology terms:
- audio processing
- automatic speech recognition
- classification
- detection and tracking
- feature analysis
- feature extraction
- feature selection
- identification
- image processing
- indexing
- language processing
- maximum likelihood
- modeling
- news story segmentation
- processing
- recognition
- retrieving
- scene fade-to-black detection
- segmentation
- segmentation work
- speaker identification
- speech recognition
- story segmentation
- topic classification
- topic detection
- topic detection and tracking
- type classification
- video analysis
- video processing
Other assigned terms:
- acoustic models
- anchor
- annotator
- asr output
- background noise
- binary feature
- broadcast news
- case
- composition
- correlation
- cue phrase
- data set
- duration
- events
- feature
- frame
- human speech
- information sources
- knowledge
- language knowledge
- likelihood
- mandarin chinese
- metadata
- methodology
- n-gram
- n-grams
- noise
- phrase
- probabilities
- probability
- segments
- signal
- silence
- speech signal
- technology
- training
- training data
- transcript
- video content
- word
- word sequences
- words
