Feature extraction and classification stage on facial expression : A review

• Main Authors: Shidiq, Muchamad Bachram, Ernawan, Ferda, Khubrani, Mousa Mohammed, Nugroho, Fajar Agung
• Format: Conference or Workshop Item
• Language: English; English
• Published: Institute of Electrical and Electronics Engineers Inc. 2022
• Subjects: QA75 Electronic computers. Computer science; QA76 Computer software; T Technology (General); TA Engineering (General). Civil engineering (General)
• Online Access: http://umpir.ump.edu.my/id/eprint/39334/1/Feature%20extraction%20and%20classification%20stage%20on%20facial%20expression_A%20review.pdf; http://umpir.ump.edu.my/id/eprint/39334/2/Feature%20extraction%20and%20classification%20stage%20on%20facial%20expression_A%20review_ABS.pdf; http://umpir.ump.edu.my/id/eprint/39334/; https://doi.org/10.1109/ICICoS56336.2022.9930545
• Institution: Universiti Malaysia Pahang Al-Sultan Abdullah

Human facial expression becomes an important technology in recent years. As information technology and networks have grown, identification and authentication have become more frequent in people's daily lives, especially using biometric technology. Human facial recognition involves face detection, feature extraction, and classification. A lot of experiments showed that there are various techniques for extracting facial features and classifying facial expressions. This paper reviews and analyze the various optimization techniques on extract feature and classification stage for human facial expression recognition. This review will compare two kinds of extract features methods and one classification method. The first technique of extracting features is the optimization technique using the K-Mean algorithm, which helps to increase recognition accuracy. The second extract feature is an optimization technique using improved Gradient Local Ternary Pattern (GLTP) which is beneficial for computational resources efficiency. Lastly, the optimization technique for image classification using a three-staged Support Vector Machine (SVM) is very helpful for increasing accuracy and eliminating error. The modified GLTP is able to obtain an accuracy of 97%.