Multi-View Frequency-Attention Alternative to CNN Frontends for Automatic Speech Recognition.
Belen AlastrueyLukas DrudeJahn HeymannSimon WieslerPublished in: INTERSPEECH (2023)
Keyphrases
- multi view
- automatic speech recognition
- speech recognition
- single view
- speech signal
- multiple views
- depth map
- hidden markov models
- d objects
- broadcast news
- conversational speech
- three dimensional
- multi view learning
- multi view clustering
- speech retrieval
- multi view stereo
- multi view reconstruction
- view synthesis
- range images
- multi view images
- co training
- surface reconstruction
- image quality
- free viewpoint
- semi supervised
- computer vision
- learning algorithm