Amirkabir University of Technology- Inventions, Discoveries and Achievements
Automatic Detection of Facial Expression in Uncontrolled Images

Clear images and colors  | Post date: 2020/12/1 | 
Researchers of Amirkabir University of Technology were able to successfully analyze facial expressions using artificial intelligence technology which can detect facial expressions independently of a person in uncontrolled images.
Dr. Hamid Sadeghi, a graduate of AUT and the project executer of "facial expression detection independent of a person in uncontrolled images" said: Facial image analysis has many applications in today's world. One of the most important analyses in the face is the recognition or identification of the emotion that the person expresses in his face (anger, hatred, fear, joy, sadness, surprise and normal), which is called facial expression. Uncontrolled images are images that are recorded outside the laboratory and in a real environment. For example, in these images, the person does not look at the camera directly necessarily and...
"One of the most important theories in this study is a metric learning method for classifying histogram features," he said. "Researchers working in the field of machine vision know well that histogram features have a lot of use in our field and the suggested method can be very useful in solving our problem."
Referring to the applications of this project, he said: "We introduced this project with a medical application (diagnosis of pain using the face). However, the methods presented to identify facial expressions in this study can be used in various applications such as building robots with the capable of detecting human facial expressions, human-computer communication, intelligent music play proportional with the individual facial expressions, and so on. "
It should be mentioned that the advisor of this project is Dr. Abolghasem Assadollah Rai, a faculty member of Amirkabir University of Technology.
Topic URL in Amirkabir University of Technology website:
http://aut.ac.ir/find.php?item=1.408.6089.en
Back to content primary page