Facial Expression Recognition System for Analysis of Facial Expression Changes when Singing

摘要

Today, music has become one aspect that cannot be separated from everyday life. Music has a lot of influence in human life. Not only on the cognitive and motor aspects but also the Emotional aspect of human beings. Everyone has different interest in music. To know the interest of someone to a music then need an application that can know the response of someone while singing a music. One of them is the Recognition of human facial expression using affective Emotion Recognition. Affectiva emotion measurement technology has many features. One of them is to recognize human facial expressions. By calculating the value of mouth movements taken from the mouth open value that in Affectiva it can be used to know someone interest or not in music. The value of this mouth motion will calculate the difference between mouth open value and then the mouth movements value will be calculated the average value and standard deviation that will be used to determine someone interest in music. The results of the experiments in 3 different categories show the results of expression recognition for the category of understanding the lyrics as many as 9 trials, the expression is in accordance with the lyrics, while the expression that is not appropriate with the lyrics is 6 trials. For experiments that do not memorize the lyrics, 11 trials are in accordance with the lyrics, while the inappropriate expressions are 4 trials. The last category is testing the category of not knowing the song, the results of 13 trials are expressed in accordance with the lyrics and 2 trials of expressions that are not in accordance with the lyrics. So that it can be concluded that the success of this application of the suitability of the singer expression with song lyrics is 73.33%.