Effectiveness of energy separation-based instantaneous frequency estimation for cochlear cepstral features for synthetic and voice-converted spoofed speech detection.
Ankur T. PatilHemant A. PatilKuldeep KhoriaPublished in: Comput. Speech Lang. (2022)
Keyphrases
- cepstral features
- voice activity detection
- text to speech
- automatic detection
- emotion recognition
- object detection
- feature extraction
- detection algorithm
- speech recognition
- false positives
- false alarms
- noisy environments
- energy consumption
- real world
- anomaly detection
- detection method
- speech signal
- scale space
- speech synthesis
- speech quality