Scientists from Nizhny Novgorod State Technical University named after R. E. Alekseev created an application capable of recognizing emotions from a photo or video. The neural network can classify emotions from video fragments, both previously filmed and received from cameras in real-time.
It is noted that this development is relevant since the application can be useful during monitoring of the driver’s condition while driving, working with video analytics systems of a “smart city” or security systems, as well as for conducting marketing research.
It is specified that for the learning process, they used a set of 35,887 images. The database was created using Google Image Search. Each image is classified into seven emotions, namely surprise, fear, happiness, anger, disgust, sadness, and neutrality, or calmness.
Ru-Main, 29.10.2020
