Cross-Informed Domain Adversarial Training for Noise-Robust Wake-Up Word Detection.
Hyungjun LimYounggwan KimHoirin KimPublished in: IEEE Signal Process. Lett. (2020)
Keyphrases
- noisy environments
- image noise
- low signal to noise ratio
- automatic detection
- partial occlusion
- detection method
- noise reduction
- imaging artifacts
- landmark detection
- detection algorithm
- co occurrence
- missing data
- greater robustness
- false alarms
- signal to noise ratio
- domain specific
- training corpus
- random noise
- noise level
- word sense disambiguation
- training samples
- false positives
- training examples
- label noise
- feature selection
- reliable detection
- keywords
- learning algorithm
- finite element analysis
- geometric distortions
- multi agent
- information extraction
- noise model
- cross domain
- domain independent