Experts of the British company - EI Technologies - have developed Xpression intelligent system, which detects changes in the human voice. Being installed on the smartphone, the program recognizes the mood of the owner and can send data to your doctor, thus preventing the development of depression.
Analysis of emotions
Xpression creators Matt Dobson and Duncan Barclay specialize in voice analysis. Their program can guess five emotional states: calmness, happiness, sadness, anger and anxiety. The application records for the day, as the change emotions, and sends the data via e-mail to you doctor.
The program works with a remote server and sends a speech fragments in length of 200 milliseconds. There data analysis happens: the voice volume, its power, change of pitch and tempo of speech, which allows the system to estimate accurately the emotional state.
Psychologist from National Health Service UK Adrian Skinner believes that society has long needed such technology.
"Being depressed people tend to fill in diaries to cease the mood. This application will help us to track the symptoms of depression every day, monitoring the patient's mood," - said Mr. Skinner.
Matt Dobson from EI Technologies told RBC daily, that the Xpression will be available to consumers in the foloowing 12 months.
"We would like that everyone would be able to take advantage of our application, as we are convinced that many people will benefit from using it. It would be great if Xpression will be used throughout the world, but we do not expect that someone will translate it, say, in Russian, before we run "- added Mr. Dobson.
Lost in Translation
The analyst of the mobile gadgets market Mikhail Fadeev said that the basis of such decisions are artificial intelligence systems, basedon the principle of the neural network.
"A similar mechanism is used in the voice input (Siri), handwriting, and in the recognition systems of printed text (OCR). The system is initially "coached" by developers in the large array of standard samples. For example, many (hundreds, thousands) people recite reference phrases. Before using this system, each person shall "teach" it to the specifics of the voice - to say the same phrase with various emotions. Later the program monitors the variations in voice and volume changes in the gaps between sounds and words, and compares it with the profile of human rights and on the basis of deviations and changes concludes mood, - says Mr. Fadeev. - For example, when a person is nervous, he is in most cases begins to speak louder and faster ".
The expert added that the application requires a very careful approach to the introduction of support for different languages: the fact that the British, Russian, Chinese, and representatives of all other nations, as a rule, speak at different speeds. In addition, people of different ages and sexes also speak with different intonations.
Комментариев нет:
Отправить комментарий