Our voice is perhaps one of the most versatile means of communication. And even if we speak different languages, you can always use universal programs for translationto understand the interlocutor. But what if the person is not yet able to speak, but still tries to communicate something? Yes, we are talking about children’s crying. It’s not just an incoherent jumble of sounds. This is an attempt to say something. At least, this concept was followed by the researchers from the United States, developing new translator of crying babies.
It is worth noting that many of dubious quality apps flooding Google Play and Apple Store, now we do not take into account. The program is a serious scientific finding.
How did the idea of creating a translator of crying babies
Actually, it’s pretty simple and the idea is not new, but one of the developers, Professor from Northern Illinois University, Lichuan Liu in the process of creating relied on the experience of pediatricians.
“Experienced nurses or pediatricians can determine why the baby is crying. Experts can do it, based on my own experience. We talked with many doctors and they confirmed that theoretically in the children’s screaming and crying it is possible to identify some patterns that you can use when creating algorithms for the translator.” says Professor Liu.
Therefore, Liu and the team decided to identify the features of children’s screams, which signal the presence of pain or discomfort These features include the difference in the intensity and frequency of crying or screaming. The team then developed an algorithm based on automatic speech recognition for detection and identification of these functions.
This “recognition algorithm language Creek” was trained on recordings of children’s screams, taken from the intensive care unit newborn. It uses a unique system that recovers the signal on the basis of incomplete data (that is, even if the system hears snatches of screaming, she can “think out” how it would look fully and make the decision). It is necessary to identify sounds produced in a noisy environment. The algorithm can identify a baby crying in the background, say, the speech of the adult or even the loud sounds of a running TV. Moreover, it is able to work in conditions when the room has a few kids, picking out the cacophony of sounds, those that can warn of danger.
This is interesting: Fish and bees learned to communicate with robots-interpreters
How accurate is the translator
Then it was time of experiments. A team of scientists has assembled skilled health workers, which gave a preview of 48 different recordings of crying babies. These same records analyzed machine. And if the people in 70% of cases “translated” lament is true, the algorithm is not much strayed away from them, “agreeing” with the doctors almost completely.
Now Professor Liu is looking forward to cooperating with companies producing assistive devices for transfer of the child’s cry. In addition, scientists continue to improve the technology for greater accuracy and expanding the number of recognized sounds. The experts also plan to add additional features such as the ability to identify and classify facial expressions of young children. This can help to give more details about emotional and physiological state of a small child.
Want more news from the world of high technology? Go to our page on Yandex.Zen. There you will discover unique materials which you can’t find on the website.