Sign language is the voice for all our fellow human beings. Even deaf people of different nationalities from all over the world can communicate with each other, since sign language has a lot in common in all languages.
But this is not enough for the full integration of the deaf in our society, since there are very few non-deaf people who know the sign language. The solution came to give the invention of a young engineer, simple as an idea, but huge because of the possibility it gives. Priyanjali Gupta is a student of the “Vellore Institute of Technology” and has created an artificial intelligence program, which has the ability to educate itself!
The rationale of the program is that it collects images of people who “speak” sign language and registers them in a database. On this basis the images are categorized based on what they mean in American Sign Language (ASL) and it automatically translates them into English.
Other attempts have been made to translate the sign language directly, but the attempts so far have been more like an ancient “systran” translation program, and the results have been at least sometimes funny, if not completely incomprehensible. The problem is that everyone makes the same letter, or word, of the sign language in a slightly different way, just as it does with the tone of voice.
The advantage of the young engineer’s program is that it constantly collects data in its database; that is, it constantly enriches all possible ways in which something can be expressed in sign language and thus become as understandable as possible, with its translation capabilities evolving constantly. Then, with the cell phone camera, as if scanning a QR code, we can read the sign language and communicate with a deaf person without knowing the sign language; impressive right?
Source:
(Συνολικές Επισκέψεις: / Total Visits: 29)
(Σημερινές Επισκέψεις: / Today's Visits: 1)