Facial Expression Recognition in Perceptual Color Space

Abstract

This paper introduces a tensor perceptual color framework (TPCF) for facial expression recognition (FER), which is based on information contained in color facial images. The TPCF enables multi-linear image analysis in different color spaces and demonstrates that color components provide additional information for robust FER. Using this framework, the components (in either RGB, YCbCr, CIELab or CIELuv space) of color images are unfolded to two-dimensional (2- D) tensors based on multi-linear algebra and tensor concepts, from which the features are extracted by Log-Gabor filters. The mutual information quotient (MIQ) method is employed for feature selection. These features are classified using a multi-class linear discriminant analysis (LDA) classifier. The effectiveness of color information on FER using low-resolution and facial expression images with illumination variations is assessed for performance evaluation. Experimental results demonstrate that color information has significant potential to improve emotion recognition performance due to the complementary characteristics of image textures. Furthermore, the perceptual color spaces (CIELab and CIELuv) are better overall for facial expression recognition than other color spaces by providing more efficient and robust performance for facial expression recognition using facial images with illumination variation.

Micro-expression Recognition Using Dynamic Textures on Tensor Independent Color Space

“…However, only gray video clips were used in [14] and [12]. Recent research shows expression recognition gets better performance in the perceptual color spaces [18]. Motivated by these researches, we propose an novel idea to use LBP-TOP on Tensor Independent Color Space (TICS) for micro-expression recognition.…”

“…To be more robust for the noise, they [17] used elastic net to make TDCS sparse and proposed Sparse Tensor Discriminant Color Space (STDCS). Lajevardi and Wu [18] treated a color facial expression image as a 3rd-order tensor and showed that the perceptual color spaces (CIELab and CIELuv) are better for facial expression recognition than other color spaces.…”

Survey on an Accurate Facial Emotion Recognition System Based on Pca

“…It concludes that by using grey-scale images we can get facial expression recognition. But we want robust recognition, that's why we have to use color images in proposed method [18].…”

Intra-Class Variation Reduction Using Training Expression Images for Sparse Representation Based Facial Expression Recognition

“…Hundred neutral face images of 100 subjects were used for determining the neutral states of faces for extracting the features of [7]. By using a random partition, following the method in [27], 1,200 face images were used for the training set and the remaining 1,200 face images were used for the test set. In the experiment with the proposed method, the training face images with sufficiently high intensities were selected and used for generating intra-class variation images.…”