How Speech Technologies Can Level the Playing Field
Closed captioning for the hearing impaired and reading assistance for the visually impaired have been available for many years, but recent new assistive technologies—as well as the continued evolution and advancement of underlying technologies—is helping an increasing number of physically challenged people use computers, the internet, and related technologies that have become mainstays of everyday life.
Assistive technologies are incorporating artificial intelligence (AI) and machine learning as well as benefiting from faster processing speeds of computers and other devices, faster internet connectivity, and in some cases the use of cloud technology, which enables more frequent updates of any underlying software.
What follows is a rundown of the innovations and vendors that are employing speech tech to help the physically challenged lead richer lives.
Visual Speech Recognition
Visual Speech Recognition (VSR) combines automated lip reading and deep neural network techniques to determine the words being spoken as well as the intent. According to Dan Miller, lead analyst and founder of Opus Research, the technology has numerous potential applications for people using smartphones, computers, tablets, and laptops, which already have the video cameras and microphones necessary to support the technology.
In his whitepaper “Introducing Visual Speech Recognition,” Miller says that smart speakers with video display are another target market as technology providers seek to improve voice-first services.
Miller also cites the ability of VSR technology to decipher speech in noisy environments, such as automobiles and crowded rooms. It can also be used in hospitals or other settings where the user may not be able to speak audibly.
Liopa’s LipRead is the first entrant into this market, though others will certainly follow if it proves successful. The mobile-based solution “captures a video of a person ‘speaking,’ which is then uploaded to Liopa’s cloud-based visual speech recognition (VSR) engine, LipRead, for processing,” according to the company. The translated phrase is then displayed on a device screen.
Because the software is on-device, it means the solution is accessible to almost anyone with a smartphone. Additionally, it’s self-tuning with a configurable vocabulary.
Liopa is targeting tracheotomy patients to begin with, but says it intends to test use in noisy, industrial settings.
Bar Code Readers
The OrCam MyEye 2 is a small, inconspicuous camera that attaches to a pair of glasses to read and relay bar codes, product logos, and text via an earpiece. The user simply needs to look at what he or she wants to read and then the attached camera provides the detail.
Sometime after the original product debuted a couple of years ago, Industries for the Blind and Visually Impaired—a manufacturer that employs many people who are blind or visually impaired—inquired about the ability to read bar codes, said Bryan Wolynski, a career optometrist and OrCam consultant.
“OCR [optical character recognition] can’t read bar codes or logos,” Wolynski says. Neither could the first edition of MyEye. Bar code reading capability was included along with other upgrades in the second edition of the device, which debuted in early 2018. “We listened to the community; this is what our user community was asking for.”
The ability for the visually impaired and other warehouse employees to quickly read bar codes is becoming more important as the growing e-commerce industry seeks to increase efficiency in moving products from storage to shipping. OrCam stores about 1 million bar codes, which it will automatically recognize and relay to the user via the earpiece. If the camera comes across a bar code that it doesn’t recognize, it can be “taught” the bar code for future use, Wolynski says.
In addition to reading bar codes, the MyEye 2 also provides improved illumination for reading in poor light conditions and a wireless battery pack for recharging.
The device can also help visually impaired employees read meeting handouts, enabling the employee to fully participate without having to go back to a scanner to read the materials, Dr. Wolynski says.
OrCam Read is a digital assistive reading device that converts text to speech.
Thanks to voice banking, digitized speech retains the essence of the original speaker
The market for assistive tech is growing, but pain points still exist
New startup utilizes 21st century AI technology to help speech-language pathologists and the language impaired.