A Wearable Speech Translation Device That Translates Voices by Easy, Hands-free Operation
ID Card Type Hands-free Speech Translation Device
Communication Barriers with an Ever Increasing Number of Foreign Patients
Recently, many foreign tourists have been visiting Japan's cities. According to the Japan National Tourism Organization (JNTO), the number of foreign visitors to Japan in June 2017 was 2,347,000, an increase of 18.2% compared to the same month in 2016.
For foreign visitors, the language barrier is something they must overcome. Could we harness the power of ICT to create a dream translation machine that instantly translates between languages, like the one that appeared in a science fiction movie a while ago?
Facilitating Smooth Provision of Medical Services with Hands-free Technology
Fujitsu Laboratories has developed an ID card type speech translation device that leaves users' hands completely free. The device, which is worn on the chest, automatically recognizes pauses in conversations and translates conversation content. This speech translation device is focused on applications in the medical field, where there is a rapidly growing need to communicate with foreign patients.
In healthcare situations, healthcare providers such as doctors and nurses often use both hands when providing diagnosis and treatment services. This compact, lightweight (*) speech translation device can be worn stress-free. Of course, it can also be used while moving.
*: 75 mm (length) x 95 mm (width) x 7 mm (height), 65 g (weight)
To design the unit structure, Fujitsu Laboratories miniaturized and optimized the sound channel configuration. We considered ease of use in healthcare situations and used miniaturization and weight reduction techniques developed by Fujitsu Connected Technologies Limited while making smartphones and other mobile phones. Fujitsu Laboratories decided to use a hanging name badge form factor that enables the healthcare provider to freely use both hands. The button icons, form, and markings facilitate intuitive operation, and the rounded shape provides a pleasant, unobtrusive impression to both the healthcare provider and patient.
The newly developed ID card type hands-free speech translation device
Speech Detection Accuracy of 95% in Healthcare Situations with Much Noise
This device employs various innovative techniques, including omnidirectional microphones, to make it compact and lightweight.
Devices that must detect the direction of the speaker, such as speech translation devices, usually employ unidirectional microphones to ensure that they pick up the speaker's voice. However, unidirectional microphones are difficult to miniaturize, which has been a major bottleneck to installing them in ID card type devices.
To address this problem, Fujitsu Laboratories focused on the "sound channel," which is the means by which the speaker's voice is transmitted to the microphone. Using an L-shaped sound channel, we successfully distinguished the target speaker's voice to be translated from sounds from other directions.
Fujitsu Laboratories adopted a high-sensitivity microphone element for the patient's direction to increase recording levels for the patient's voice. By combining the L-shaped sound channel with noise suppression technology, this device has achieved a high speech detection accuracy of 95% even in healthcare environments with much noise coming from air conditioners and diagnostic devices.
Another feature is that actual translation is performed on a server (**). Although it supports only English and Chinese at the moment, support for other languages may be added by upgrading the server. Since this system can easily support multiple languages without making any changes to the device itself, it has the potential to be applied in a variety of fields beyond medicine.
**: This uses multi-language speech translation technology of the National Institute of Information and Communications Technology (NICT).
Offering "Omotenashi" (Hospitality) Services through Multilingual Translation
This wearable speech translation device can be connected to cloud-based translation programs; you can prepare programs in the cloud that serve different purposes depending on your requirements.
This speech translation device is particularly suited to providing omotenashi (hospitality) services to foreign tourists visiting Japan at international sporting events planned to be held in the near future. We will continue to do our best to evolve its function and performance in the future.