A Survey on Various Techniques of Human Emotion Recognition and its Applications
- February 24, 2019
- Posted by: RSIS
- Category: Computer Science and Engineering
International Journal of Research and Scientific Innovation (IJRSI) | Volume VI, Issue II, February 2019 | ISSN 2321–2705
Anjana C R
Department of Computer Science, Jawaharlal College of Engineering and Technology, Kerala, India
Abstract— Humans show a variety of emotions according to the situation and identifying these emotions can be helpful in efficient communication. In the case of human- machine interaction, identifying the human’s emotion correctly is very important. In the current situation where machines act as an inevitable part of human life as an instructor, a helper, or even a companion, it would be helpful if the machines could understand the emotions of the human and act accordingly. Human emotions can be recognized from the tone of voice, facial expressions (which may be from image or from video), body language and using EEG (Electroencephalography). EEG is a technique that uses electrodes attached to the scalp to identify the tiny electrical charges that result from the activity of the brain cells for emotion detection . But it is not always possible to wear the electrodes in order to communicate with a machine. So, more focus is given to the other techniques of emotion recognition. Emotion recognition from facial images is done by using image processing techniques. Various types of neural networks can be employed for identifying various expressions from faces in the images. Acoustic data can be collected and used for identifying emotions from sound input. The body language of a person can also convey some information about the emotional state of the individual.
Keywords— Neural Network, Mel Spectrogram, Chromagram, Electroencephalography
Emotion recognition is always important for meaningful communication. Facial emotions are important cues for non-verbal communication. This is why humans can recognize emotions accurately. The basic facial expressions convey discrete and specific emotions: anger, happiness, surprise, fear, disgust, and sadness. The face communicates a lot of information such as the emotional state of the person and their identity.
For facial emotion recognition, we have to identify the face from the given image. The algorithm proposed by Viola and Jones is the most commonly used method for face identification. It used the textures that are common to all faces such as the lighter areas above and below the eye, the edge features and the darker areas to the left and right of the nose to locate the face.