Phonetic Segmentation Using KALDI and Reduced Pronunciation Detection in Causal Czech Speech.
Zdenek PatcPetr MizeraPetr PollákPublished in: TSD (2015)
Keyphrases
- speech recognition
- noisy environments
- speaker independent
- hidden markov models
- speech signal
- automatic speech recognition
- speech retrieval
- speech recognizer
- language model
- level set
- speech synthesis
- spontaneous speech
- image segmentation
- speech segments
- acoustic models
- segmentation algorithm
- speaker identification
- object segmentation
- bounding box
- voice activity detection
- multiscale
- pattern recognition
- segmentation method
- cell segmentation
- detection method
- object detection
- reliable detection
- speech recognition systems
- low signal to noise ratio
- complex background
- region growing
- spoken term detection
- spoken document retrieval
- human machine interaction
- language independent
- causal relationships
- causal models
- automatically generated
- detection algorithm
- medical images
- word segmentation
- vocal tract
- digit recognition
- cross language
- edge detection
- information retrieval