Sony's New Lip-Reading Technology Could Boost Accessibility—or Invade Privacy

 

facial-recognition software can discover faces in a crowd, but how approximately picking up conversations with out the assist of close by microphones? sony's visual speech enablement does just that, the usage of digital camera sensors and ai for augmented lip studying in any surroundings.

https://www.pcmag.com/how-to/how-to-transfer-facebook-photos-videos-to-google-photos

mark hanson, sony’s vp of product generation and innovation, gave a restrained evaluate of the technology at some point of a ces keynote. it is a brand new use case for sony's wise imaginative and prescient photograph sensor and uses ai to isolate a consumer’s lips after which translates their actions into words, impartial of any history or foreground noise. in truth, it requires no microphone in anyway. the gap among the sensor and person is sort of inconsequential and it can paintings over many feet, virtually by the use of a better-resolution sensor, hanson informed us ultimate week.

https://www.linkedin.com/company/nextdottech

sony initially plans to marketplace the generation for a handful of use cases, which include factory automation, kiosks, and voice-enabled atms. visual speech enablement is optimized to be used on computers, even though purchaser-facing versions of the characteristic should roll out on mobile hardware inside the destiny, consistent with hanson, who sees it as an assistive generation, now not a surveillance tool. it may enhance car-generated captions, as an instance, or lessen the need for a relay operator or automated speech-popularity intermediary that calls for a stable records connection and minimal historical past noise.

https://www.pinterest.com/shorthandapp/health-and-medicine/

however for all of its capacity for proper, there’s additionally the opportunity it is able to be misused. hanson says the technology simplest captures lips, not faces, so no consumer-identifiable facts is retained. what remains unaddressed is the opportunity of combining visible speech enablement with different technologies, lots of which use cameras and will incorporate sony's ai-better sensors. if visible speech enablement have been to sit along a facial-reputation camera, the facts will be aggregated and undo sony's integrated privateness protections.

https://www.nytimes.com/2020/07/01/technology/personaltech/make-your-tech-last-longer.html