referenceArchiveUUID: "73c5b458-fda6-44b7-865c-07afb79f0c95"
fullTitle: "An Introduction to Silent Speech Interfaces"
authors: ["João Freitas","António Teixeira","Miguel Sales Dias","Samuel Silva"]
A short (~100p) book-length overview of 2017-era progress on Silent speech interface technology, across many modalities. Includes an overview of how human speech production works, from neurophysiology through to facial actuation.
This book was a nice way to get an overview of the different modalities, but its portrayal of their relative progress seems already outdated / untrustworthy: the deep learning revolution was only just beginning to enter this research space when this book was published. For instance, this book doesn’t (and couldn’t) include AlterEgo (2018, SEMG + DL) or Prajwal et al’s 2022 transformer-based Visual speech recognition work, which seem like an enormous leap over the rates presented.
At the end of the book, I found myself interested in more coverage of “non-audible murmur” microphones (see page 43)—they might be quiet enough for Tools for thought purposes. See e.g. Fukumoto - 2018 - SilentVoice