A new neural network can distinguish between individual people and objects in a video, as well as speed them up or slow them down
For this, the Google and Oxford teamThe university divided each video frame into separate layers and taught AI to identify people or objects in them. This neural network focuses on things on each layer, concentrating on their movements, then it can separate each object and change their movement.
Scientists have determined the upper limit of the speed of sound
Previously, Google came up with a model that couldread sign language during video calls. AI can detect who's actively speaking, but ignores the other person if they're just moving their hands or head. A new development from Google researchers can do this with great efficiency and low latency. So far, the researchers note that sign language detection leads to delays or deterioration in video quality, but this problem can be solved, and the model itself remains lightweight and reliable.
The system first runs the video through the model undercalled PoseNet, which assesses the position of the body and limbs in each frame. Simplified visual information is sent to a model trained to position data from videos of people using sign language and compares the image to how people usually display certain words.
Read also
On day 3 of illness, most COVID-19 patients lose their sense of smell and often suffer from a runny nose
See the closest shots of the Sun's surface
The Doomsday glacier turned out to be more dangerous than scientists thought. We tell the main thing