On-screen keyboards are everyday life, but somehow impractical - on the one hand, they cover a large part of the screen, especially on smartphones, and on the other hand, ten-finger typing usually doesn't really work.
Researchers at the Korea Advanced Institute of Science and Technology ( KAIST ) have now developed a completely virtual soft keyboard that is intended to help. The “I-Keyboard” relies on deep learning to recognize what the user is typing – no matter where and how on the screen.
Invisible keyboard
“The proposed I-Keyboard is invisible, maximizing the usefulness of the screen on mobile devices. Users can view the content of an application in full-screen mode and type freely at the same time.”
write researchers Ue-Hwan Kim, Sahng-Min Yoo and Jong-Hwan Kim in a paper published in advance on “arXiv”. The concept is therefore clearly aimed at users who have actually mastered the ten-finger system. Of course, it promises particularly easy and flexible use at any position on the screen and at any typing angle.
[mk_ad]
There is no predefined keyboard layout; the shape and size of the virtual keys are also variable. According to the team, the I-Keyboard instead uses a deep learning-based decoding algorithm to recognize what the user is typing. In principle, the AI system continuously adapts the virtual keyboard to what the user imagines. The I-Keyboard can therefore immediately compensate for the fact that a user's fingers drift away from their initial position over time due to the lack of tactile feedback of an on-screen keyboard.
Performance increase
According to the researchers, experiments have shown that the approach works and appears to be superior to previous soft keyboards for ten-finger use. The accuracy of the detection is therefore 95.84 percent, slightly higher than previous solutions. The effects are greater when it comes to writing speed. With the I-Keyboard, users can achieve 45.57 words per minute (wpm), which is an increase of almost a fifth compared to the comparison basis. However, it doesn't come close to a real keyboard, but in comparison costs around a ninth of a wpm.
To the paper “ I-Keyboard: Fully Imaginary Keyboard on TouchDevices Empowered by Deep Neural Decoder ”
Source: press text
Article image: Shutterstock / By Peshkova
Notes:
1) This content reflects the current state of affairs at the time of publication. The reproduction of individual images, screenshots, embeds or video sequences serves to discuss the topic. 2) Individual contributions were created through the use of machine assistance and were carefully checked by the Mimikama editorial team before publication. ( Reason )