Open AccessBasic Emotion Recogniton using Automatic Facial Expression Analysis Software
Author(s) -
Vivi Triyanti,
Yassierli Yassierli,
Hardianto Iridiastadi
Publication year - 2019
Publication title -
jurnal optimasi sistem industri
Language(s) - English
Resource type - Journals
eISSN - 2442-8795
pISSN - 2088-4842
DOI - 10.25077/josi.v18.n1.p55-64.2019
Subject(s) - surprise , facial expression , emotion classification , anger , expression (computer science) , psychology , subject (documents) , cognitive psychology , emotional expression , software , computer science , social psychology , communication , library science , programming language
Facial expression was proven to show a person's emotions, including 6 basic human emotions, namely happy, sad, surprise, disgusted, angry, and fear. Currently, the recognition of basic emotions is applied using the automatic facial expression analysis software. In fact, not all emotions are performed with the same expressions. This study aims to analyze whether the six basic human emotions can be recognized by software. Ten subjects were asked to spontaneously show the expressions of the six basic emotions sequentially. Subjects are not given instructions on how the standard expressions of each of the basic emotions are. The results show that only happy expressions can be consistently identified clearly by the software, while sad expressions are almost unrecognizable. On the other hand surprise expressions tend to be recognized as mixed emotions of surprise and happy. There are two emotions that are difficult to express by the subject, namely fear and anger. The subject interpretation of these two emotions varies widely and tends to be unrecognizable by software. The conclusion of this study is the way a person shows his emotions varies. Although there are some similarities in expression, it cannot be proven that all expressions of basic human emotions can be generalized. Further implication of this research needs further discussion.