Academia.eduAcademia.edu

Figure 4 – uploaded by Nouha Khédiri

See full PDFdownloadDownload figure
Fig. 4. Eyes, Eyebrows and Mouth Detection [6]

Figure 4 Eyes, Eyebrows and Mouth Detection [6]

Related Figures (4)

Random Forest based classification algorithm. Another work by Razek and Frasson in [10] describe a text-based emotion recognition approach using Dominant Meaning Technique (DMT). This technique is based on the meaning of the word and allows to build the dominant meaning tree using ISEAR (International Survey on Emotion Antecedents and Reactions) dataset. Emotion Detection Algorithm was created to classify seven emotions which are joy, fear, anger, sadness, disgust, shame and guilt. tracking. The classification method used in [6] is simple and based on the variation of some distances from the neutral face. In the work done by Yang et al. [7], a computer vision system that automatically analyze the learning status of their students during the learning has been implemented to enhance an online Virtual English Classroom called WEC3D. Facia features are extrated using AdaBoost classifier in a feature vector form, then, each feature vector sequence is recognized and trained using Hidden Markov Model (HMM). Finally, an expression vector, that contains six expressions, is analyzed and a Gaussian Mixture Model (GMM) is applied to evaluate three learning scores such as Understanding, Interaction, and Consciousness [6].
Random Forest based classification algorithm. Another work by Razek and Frasson in [10] describe a text-based emotion recognition approach using Dominant Meaning Technique (DMT). This technique is based on the meaning of the word and allows to build the dominant meaning tree using ISEAR (International Survey on Emotion Antecedents and Reactions) dataset. Emotion Detection Algorithm was created to classify seven emotions which are joy, fear, anger, sadness, disgust, shame and guilt. tracking. The classification method used in [6] is simple and based on the variation of some distances from the neutral face. In the work done by Yang et al. [7], a computer vision system that automatically analyze the learning status of their students during the learning has been implemented to enhance an online Virtual English Classroom called WEC3D. Facia features are extrated using AdaBoost classifier in a feature vector form, then, each feature vector sequence is recognized and trained using Hidden Markov Model (HMM). Finally, an expression vector, that contains six expressions, is analyzed and a Gaussian Mixture Model (GMM) is applied to evaluate three learning scores such as Understanding, Interaction, and Consciousness [6].
Fig. 1. Emotional Recognition Process
Fig. 1. Emotional Recognition Process
Fig. 2. Methodology of Facial Recognition in EMASPEL [6]
Fig. 2. Methodology of Facial Recognition in EMASPEL [6]
Fig. 3. Facial Detection [6] For feature points extraction (cf. Fig.2), there are a three Fig.3) which are: (1) Skin Color [21] that needs firstly a preprocessing, (2) SMA face detection [22] that contains eye location, mouth location and face location, and (3) Real time face detection [23] based on AdaBoost which detects automatically the face of the learner under several constraints like distance, luminosity and varied background. It can also detect more faces in one image.
Fig. 3. Facial Detection [6] For feature points extraction (cf. Fig.2), there are a three Fig.3) which are: (1) Skin Color [21] that needs firstly a preprocessing, (2) SMA face detection [22] that contains eye location, mouth location and face location, and (3) Real time face detection [23] based on AdaBoost which detects automatically the face of the learner under several constraints like distance, luminosity and varied background. It can also detect more faces in one image.
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved