Video indexing by emotions

The Australian company Emotiv Systems , which since 2003 has been developing devices for organizing interfaces and controlling computers using brain technology based on electroencephalography (EEG), in 2009 released the EPOC gaming device. It is a headset with fixed electrodes that read the human brain impulses, allowing you to control the gameplay, just thinking about the necessary action. There are several games on the program's website where you can try the game controller in action, as well as a photo viewer in which all control is performed using thoughts.

And this unusual device seems to have found a new use.

This video demonstrates the work of the EmoRate program, which uses the Emotiv 14-electrode headset to record emotions by reading the facial expressions of your face. While a person is watching videos on YouTube, EmoRate records his emotions, after which he indexes the watched videos, assigning them to recorded emotions. After indexing, you can navigate through the clips that cause the emotions that you want to receive.

On the EmoRate page, its creator tells how the Emotiv SDK was used to create the program and how the system was trained.
')
The application demonstrates the power of affective computing, a technology that allows computers to identify and respond to human emotions and drastic changes in our behavior in our daily interaction with computers.

The device, developed by Emotiv, made it possible to remove the last barrier to the development of affective computing. Despite the fact that the human consciousness is capable of experiencing many different emotions, four of the most powerful were taken as the basis: happiness, sadness, anger, and fear. So far, the Emotive learning algorithm of the device allows you to accurately determine only these emotions, but in the future, as the algorithms improve, this list should be updated.

When you experience certain emotions, you involuntarily use the various muscles of your face that the program analyzes. To use technology, the program needs to be trained to recognize exactly your emotions.

The number of possible applications of the technology is already quite large. The author suggests the term Emoprofil, a profile that contains data describing a person’s emotional reactions to various stimuli. With a sufficient base of recognized emotions of different people, having found patterns, profiles can be applied in various situations. For example, to determine whether a person loves another or not, whether people are compatible with each other and how they will interact within the group.

Emotions allow you to implement new recommendation systems that can most accurately select movies, books, games, etc., under your emotional profile, which the existing recommendation systems on various sites cannot afford.

Imagine the search engine, where what you feel at the time of sending the request will play a key role in the search results. For example, a scared cancer patient needs help and he will see a list of hospitals and support groups, while a curious post-graduate student will receive a list of recent research on the topic of dealing with the ailment and treatment methods.

Dating services based on emotional profiles will definitely replace traditional ones, as people will know in advance if they are suitable for each other and whether to start or continue communication. And together with geolocation will allow to meet people who will be interesting and pleasant for you.

Emotions will allow to refuse captcha when entering passwords for access to your personal web services. Your emotional profile allows the service to determine what it is you are trying to access.

The spread of emotional profiles will entail the emergence of new websites, games and other entertainment that will use your emotions. And the number of applications of this technology can go on and on.

On the project website, you can follow the development of the program and receive the latest news from the field of affective computing and brain-machine interfaces.

Based on slashdot