Science in Information Technology Letters Vol. 2., No. 1, May 2021, pp. 43-53 ISSN 2722-4139 http://pubs2.ascee.org/index.php/sitech http://dx.doi.org/10.31763/sitech.v1i1.1 sitech@ascee.org Image processing for student emotion monitoring based on fisherface method A H Pratomo a,1,* , M Y Florestyanto a , Y I Sania a , B Ihsan b , H H Triharminto b a Jurusan Informatika, Fakultas Teknik Industri, Universitas Pembangunan Nasional Veteran Yogyakarta b Department of Electrical Engineering, Indonesian Air Force Academy 1 awang@upnyk.ac.id * Corresponding Author 1. Introduction Academic emotion is one of the success factors for students learning in class. The emotions are recognized by looking at student's expressions in the class where change over time [1]. Emotional monitoring is done by detecting students' faces and then identifying them. After the identification is made, the process continues with emotion detection using a webcam as a sensor device for capturing the image. The image is then identified based on image extraction, and facial landmarks are searched to detect emotions. Conducted emotion detection using a face-based template based on fuzzy classifier [2]. The other method for emotion detection is based on facial landmarks to determine changes in facial expressions. Face detection is performed to determine the face object [3]. One way to detect faces is to detect skin color. Research related to skin color detection was carried out by [4] using RGB, HSV, and YCbCr color comparisons to distinguish skin color from the background. The development of the method can be used for face detection. [5] used the skin color detection method combined ARTICLE INFO ABSTRACT Article history Received March 12, 2021 Revised April 20, 2021 Accepted May 15, 2021 Monitoring academic emotion is an activity to provide information from students' academic emotions in the class continuously. Some research in the image processing field had done for face recognition but had not been many studies on image processing to detect student emotions. This paper aims to determine the percentage of facial recognition with fisherface and academic emotional recognition by monitoring changes in students' facial expressions using facial landmarks in various distances, camera angles, light, and attributes used on objects. The proposed method uses facial image extraction based on fisherface method for presence. Furthermore, face identification will be made with Euclidean distance by finding the smallest length of training data with test data. Emotion detection is done by facial landmarks and mathematical calculations to detect drowsiness, focus, and not focus on the face. Restful web service is used as a communication architecture to integrate data. The success rate of applications with the fisherface method obtains 96% percent accuracy of face recognition. Meanwhile, facial landmarks and mathematical calculations are used to detect emo tions, with 84 %. This is an open access article under the CC–BY-SA license. Keywords Image processing Facial recognition Fisherface