Limits...
Robust facial expression recognition via compressive sensing.

Zhang S, Zhao X, Lei B - Sensors (Basel) (2012)

Bottom Line: The CS theory is used to construct a sparse representation classifier (SRC).The effectiveness and robustness of the SRC method is investigated on clean and occluded facial expression images.Three typical facial features, i.e., the raw pixels, Gabor wavelets representation and local binary patterns (LBP), are extracted to evaluate the performance of the SRC method.

View Article: PubMed Central - PubMed

Affiliation: School of Physics and Electronic Engineering, Taizhou University, Taizhou 318000, China. tzczsq@163.com

ABSTRACT
Recently, compressive sensing (CS) has attracted increasing attention in the areas of signal processing, computer vision and pattern recognition. In this paper, a new method based on the CS theory is presented for robust facial expression recognition. The CS theory is used to construct a sparse representation classifier (SRC). The effectiveness and robustness of the SRC method is investigated on clean and occluded facial expression images. Three typical facial features, i.e., the raw pixels, Gabor wavelets representation and local binary patterns (LBP), are extracted to evaluate the performance of the SRC method. Compared with the nearest neighbor (NN), linear support vector machines (SVM) and the nearest subspace (NS), experimental results on the popular Cohn-Kanade facial expression database demonstrate that the SRC method obtains better performance and stronger robustness to corruption and occlusion on robust facial expression recognition tasks.

No MeSH data available.


Related in: MedlinePlus

Recognition accuracy under different percentage occluded.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC3376615&req=5

f8-sensors-12-03747: Recognition accuracy under different percentage occluded.

Mentions: We next investigate the robustness of SRC to the random block occlusion. We simulate this situation under different percentage occluded, from 0% to 50%, by replacing a randomly located square block of each test image with an unrelated image of a baboon, as shown in Figure 7(a). Note that, the location of occlusion is randomly chosen for each image and is unknown to the algorithm. Figure 7 shows an example of a 30% occluded face image. To the human eye, beyond 30% occlusion, the entire facial regions have been almost completely occluded. In this case, it’s a difficult recognition task even for humans. Figure 8 gives the recognition performance of SRC and its three competitors, as a function of the percentage occluded from 0% to 50%. As illustrated in Figure 8, we can see that the recognition accuracy of SRC significantly exceeds that of other used methods at various levels of occlusion. This demonstrates SRC achieves a higher level of robustness to the random block occlusion in comparison with the other used methods.


Robust facial expression recognition via compressive sensing.

Zhang S, Zhao X, Lei B - Sensors (Basel) (2012)

Recognition accuracy under different percentage occluded.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC3376615&req=5

f8-sensors-12-03747: Recognition accuracy under different percentage occluded.
Mentions: We next investigate the robustness of SRC to the random block occlusion. We simulate this situation under different percentage occluded, from 0% to 50%, by replacing a randomly located square block of each test image with an unrelated image of a baboon, as shown in Figure 7(a). Note that, the location of occlusion is randomly chosen for each image and is unknown to the algorithm. Figure 7 shows an example of a 30% occluded face image. To the human eye, beyond 30% occlusion, the entire facial regions have been almost completely occluded. In this case, it’s a difficult recognition task even for humans. Figure 8 gives the recognition performance of SRC and its three competitors, as a function of the percentage occluded from 0% to 50%. As illustrated in Figure 8, we can see that the recognition accuracy of SRC significantly exceeds that of other used methods at various levels of occlusion. This demonstrates SRC achieves a higher level of robustness to the random block occlusion in comparison with the other used methods.

Bottom Line: The CS theory is used to construct a sparse representation classifier (SRC).The effectiveness and robustness of the SRC method is investigated on clean and occluded facial expression images.Three typical facial features, i.e., the raw pixels, Gabor wavelets representation and local binary patterns (LBP), are extracted to evaluate the performance of the SRC method.

View Article: PubMed Central - PubMed

Affiliation: School of Physics and Electronic Engineering, Taizhou University, Taizhou 318000, China. tzczsq@163.com

ABSTRACT
Recently, compressive sensing (CS) has attracted increasing attention in the areas of signal processing, computer vision and pattern recognition. In this paper, a new method based on the CS theory is presented for robust facial expression recognition. The CS theory is used to construct a sparse representation classifier (SRC). The effectiveness and robustness of the SRC method is investigated on clean and occluded facial expression images. Three typical facial features, i.e., the raw pixels, Gabor wavelets representation and local binary patterns (LBP), are extracted to evaluate the performance of the SRC method. Compared with the nearest neighbor (NN), linear support vector machines (SVM) and the nearest subspace (NS), experimental results on the popular Cohn-Kanade facial expression database demonstrate that the SRC method obtains better performance and stronger robustness to corruption and occlusion on robust facial expression recognition tasks.

No MeSH data available.


Related in: MedlinePlus