Future Speech Interfaces with Sensors and Machine Intelligence

• Drugi avtorji: Denby, Bruce (Editor), Gábor Csapó, Tamás (Editor), Wand, Michael (Editor)
• Format: Elektronski Book Chapter
• Jezik: angleščina
• Izdano: Basel MDPI - Multidisciplinary Digital Publishing Institute 2023
• Teme: Technology: general issues; History of engineering & technology; neural machine translation (NMT); transformer; Arabic dialects; modern standard Arabic; subword units; multi-head attention; shared vocabulary; self-attention; 3D densely connected CNN; 3D multi-layer feature fusion CNN; convolutional neural network; deep learning; lipreading; speech recognition; visual speech recognition; silent speech; continuous-wave radar; European Portuguese; machine learning; multimodal speech; lip reading; ultrasound tongue imaging; pose estimation; speech kinematics; keypoints; landmarks; audio-visual speech recognition; lip-reading; application programming interface; multi-modal interaction; deep neural networks; multi-view VSR; attention mechanism; spatial attention module; local self-attention; connectionist temporal classification; text-to-lip; speech synthesis; text-to-speech; speech-to-lip; zero-shot adaptation; generative models; artificial intelligence; objective measures; hybrid models; end-to-end recognition; reliability measures; decision fusion net; articulation-to-speech synthesis; silent speech interface; speaker adaption; voice conversion; audiovisual speech recognition; multimodal interaction; edutainment; virtual aquarium; speech processing; ultrasound imaging; silent speech interfaces; speech sensors
• Online dostop: DOAB: download the publication; DOAB: description of the publication

Speech is the most spontaneous and natural means of communication, as well as the preferred modality for interacting with mobile or fixed electronic devices, but speech in-terfaces have drawbacks, such as a lack of user privacy; non-inclusivity for certain users; poor robustness in noisy conditions;...