Facial expression recognition using feature based techniques and model based techniques: A survey

Abstract

Facial expression is a way of non-verbal communication. A person depicts his/her feelings through facial expressions. In computer systems facial expressions help in verification, identification and authentication. One popular use of facial expression recognition is automatic feedback capture from customers upon reacting to a particular product. Effective recognition technology is in high demand by the common users of today's gadgets and technologies. Facial expression recognition technique is broadly classified into two techniques: Feature based techniques and Model based techniques. The key contribution of this article is that we have analyzed latest state of the art techniques in Feature based techniques and Model based techniques. These techniques are analyzed using various standard public face databases: GEMEP-FERA, BU-3DFE, CK+, Bosphorous, MMI, JAFFE, LFW, FERET, CMU-PIE, Georgia tech, AR, eNTERFACE 05 and FRGC. From our analysis we found that for Feature based Curvelet approach performed on FRGCv2 database gave an excellent 97.83% recognition rate and Model based textured 3D video technique performed on BU-4DFE database gave an excellent 94.34 % recognition rate.

Facial Expression Recognition Complications with the Stages of Face Detection and Recognition

“…It characterizes the features into six layers there by expands the extraction times and calculation of the related features. Bish was Mishra et.al [42] explained a Facial expression was a method for nonverbal correspondence. A man delineates human emotions through facial emotions.…”

“…Among these multitask cascaded CNNs-based framework detects the faces with better accuracy than other methods.Also it predicts face and landmark location in a coarse-to-fine manner. [42], landmark detection method by CLM model based on DRMF method [44] to recognize the features of faces. Among these (DNN) architecture for automated FER identifies features with better accuracy than the other methods.…”

3D Approaches and Challenges in Facial Expression Recognition Algorithms—A Literature Review

“…The study of facial expressions and emotions is a topic that has interested in the research community for several years. In the absence of a uniquely comprehensive and widely accepted theory, multiple approaches for facial expression recognition have emerged, and extensive and detailed surveys of FER focusing on traditional approaches research are given in [4][5][6][7]; the reader is referred to [8,9] for 3D and [10,11] for 2D earlier approaches. Recently, deep learning-based FER approaches emerged, and have been surveyed in [1,[12][13][14].…”

“…Denotes the database, or a portion of it, used for the study; 2 indicates the use of 2D data associated, and 3 the use of temporal information (3D image sequences) 4. The presence, number, and type of facial landmarks: manual or automatic 5.…”

“…Denotes the type of data, 2 indicates the number of samples and, in brackets, the number of subjects. Presence and number of 3 occlusions of eyes and mouth,4 various poses of the head acquired for each subject in a systematic way, and 5 facial landmarks.…”