Skip to main content
Glite Dictionary
👄 To lipread 💻 To analyze speech
to speech-read

to speech-read

9.3
To use technology to identify the words a person says from their lip and face movements
  • verb
  • /spiːtʃ riːd/
  • Specialized
translation icon : leer los labios
  • The new app speech-reads speakers in real time, creating captions when the microphone is off.

Examples

  • Researchers are speech-reading video to improve accuracy of lip-reading models for noisy audio conditions.

  • The security team used software that could speech-read conversations from the silent CCTV footage.

  • The software can speech-read in real-time.

  • How well does this program speech-read?

  • The software can speech-read the speaker's words by analyzing their lip movements during the presentation.

  • Using advanced algorithms, the system speech-reads the audio-visual data to accurately transcribe the lecture.

  • The new technology speech-reads from videos, making it easier for those with hearing impairments to follow conversations.

Synonyms

to read lips
vsto speech-read
11 5.7

To understand what someone says by watching their lips

is done automatically by computer vision rather than by a person watching lips
to lip-read
vsto speech-read
4 7.2

Understand what someone says by looking at their mouth

is performed by machines to reconstruct speech from lip and face visuals
to lipread
vsto speech-read
8.4

To understand what someone says by watching their lips

is an automatic computer process using lip and face movements to reconstruct speech

How Accurate

to lip-read
4 7.2
to speech-read
  • Specialized
9.3

Surface Forms

speech-read infinitive
speech-read past
speech-read past participle
speech-reading present participle
speech-reads third person singular

Morphology

speech-read = speech + read (transparent) = speech + read

Compound formation 'speech' + 'read' equally describes the computational task of visually recognizing spoken language; components map straightforwardly to the meaning.

Etymology

Speech-read in computing keeps the same idea: a program reads speech ('spoken words') from the picture of a face, like a person reading a book with their eyes. So, to speech-read in computers is to use the mouth and face movements in a video to figure out what was said.