Investigation of nonlinear feature extraction techniques for facial emotion recognition/
The main objective of this thesis is to investigate ways to achieve robust and performing FER by exploring a number of new features based on nonlinear feature extraction techniques. The objectives of this research are to develop nonlinear feature extraction techniques for facial emotion recognition...
Saved in:
| Main Author: | |
|---|---|
| Corporate Author: | |
| Format: | Thesis Book |
| Language: | English |
| Published: |
Perlis, Malaysia
School of Mechatronic Engineering
2016
|
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | The main objective of this thesis is to investigate ways to achieve robust and performing FER by exploring a number of new features based on nonlinear feature extraction techniques. The objectives of this research are to develop nonlinear feature extraction techniques for facial emotion recognition using higher order spectra (HOS) and empirical mode decomposition (EMD) techniques, independently. The bispectral and intrinsic mode functions (IMF) features which describe the characteristic of distinct facial emotion are adopted as intrinsic features; to examine different dimensionality reduction techniques such as Linear Discriminant Analysis (LDA), Local Fisher Discriminant Analysis (LFDA) and Kernel Local Fisher Discriminant Analysis (KLFDA) on the bispectral and IMF features; to analyze the performance of reduced features using k-NN, SVM, and ELM-RBF classifiers; to evaluate and validate the proposed method in classifying seven facial emotions (anger, disgust, fear, happiness, neutral, sadness and surprise) with original, partially occluded and noise-corrupted images. |
|---|---|
| Physical Description: | xix, 176 pages colour illustration 30 cm. |
| Bibliography: | Includes bibliographical references. |
| ISBN: | X210000000779 |