Abstract: Handcrafted audio descriptors and learned deep representations each bring distinct strengths and inherent limitations to speech emotion recognition (SER). Traditional handcrafted features ...
Abstract: Noise robustness is critical when applying automatic speech recognition (ASR) in real-world scenarios. One solution involves using speech enhancement (SE) models as the front end of ASR.
Some results have been hidden because they may be inaccessible to you
Show inaccessible results