Facial Expression Recognition Based on Local Fourier Coefficients and Facial Fourier Descriptors

Abstract

The recent boom of mass media communication (such as social media and mobiles) has boosted more applications of automatic facial expression recognition (FER). Thus, human facial expressions have to be encoded and recognized through digital devices. However, this process has to be done under recurrent problems of image illumination changes and partial occlusions. Therefore, in this paper, we propose a fully automated FER system based on Local Fourier Coefficients and Facial Fourier Descriptors. The combined power of appearance and geometric features is used for describing the specific facial regions of eyes-eyebrows, nose and mouth. All based on the attributes of the Fourier Transform and Support Vector Machines. Hence, our proposal overcomes FER problems such as illumination changes, partial occlusion, image rotation, redundancy and dimensionality reduction. Several tests were performed in order to demonstrate the efficiency of our proposal, which were evaluated using three standard databases: CK+, MUG and TFEID. In addition, evaluation results showed that the average recognition rate of each database reaches higher performance than most of the state-of-the-art techniques surveyed in this paper.

Multicultural Facial Expression Recognition Based on Differences of Western-Caucasian and East-Asian Facial Expressions of Emotions

“…The FER system used for cross-cultural evaluation is based on the hybrid method originally proposed in [8] which was enhanced by the application of the Fast Fourier Transform (FFT), as detailed in [7]. The system framework follows the steps of face detection, facial region segmentation, feature extraction and classification.…”

“…Basically, LFC and FFD are based on the application of the FFT and the calculation of individual eigenspaces for each fa- cial part, which are used to obtain independent feature vectors of appearance and geometric features, respectively. Hybrid feature vectors are obtained by concatenating independent feature vectors, which in turn are projected into a final region-based eigenspace (for details of this method, please refer to [7]). Figure 4 illustrates the feature extraction process of the mouth region, where Y Y Y vec represents the hybrid feature vector, which is a projection of the concatenated H H H vec vector into the Es Es Es reg mouth Eigenspace.…”

“…The main difference is that, in this paper, the analysis was applied to a more generalized database (more than 5 times bigger than those of [8] and [9]) which includes more than 1,000 expressive faces from five standard databases (CK+, MUG, JAFFE, JACFE and TFEID). Moreover, the feature extraction process employed for the crosscultural FER analysis is fully automatic as proposed in [7].…”

Multiparameter Space Decision Voting and Fusion Features for Facial Expression Recognition

“…To elaborately extract facial expression features, some studies have divided facial images into nonoverlapping blocks [ 20 – 22 ]. Because expression features reflect the changes of the direction, edge, and intensity of the texture of an image, extracting features in regions of interest (ROIs) is a common practice [ 23 – 25 ]. Most ROIs are eye, mouth, and eyebrow regions, which are fixed in a set of sizes.…”

Merged Local Neighborhood Difference Pattern for Facial Expression Recognition

“…Even though the improvement in the recognition rate is high using deep features, the controversy among manually extracted features and deep features is indeed present. Recently, the hand-crafted feature introduced by Benitez-Garcia et al [20] is capable of attaining a higher recognition level than a deep learning model. This indicates that the handcrafted features and domain-specific familiarity are still successful and favorable in computer vision-based classification.…”