Haar Features for FACS AU Recognition

Abstract:Facial expression has lots of applications in human-computer interaction. Although feature extraction and selection have been well studied, the specificity of each expression variation is not fully explored in state-of-the-art works. In this work, the problem of multiclass expression recognition is converted into triplet-wise expression recognition. For each expression triplet, a new feature optimization model based on Action Unit (AU) weighting and patch weight optimization is proposed to represent the specificity of the expression triplet. Sparse representation based approach is then proposed to detect the active AUs of testing sample for better generalization. The algorithm achieved competitive accuracies of 89.67% and 94.09% for Jaffe and CK+ databases, respectively. Better cross-database performance has also been observed.

show abstract

“…NE , n = 10 is the number of patches set to reduce the influence of abnormal weight components by sparse representation (19).…”

Section: Active Au Detectionmentioning

confidence: 99%

Active AU Based Patch Weighting for Facial Expression Recognition

Xie

Yang

Lai

2017

Preprint

View full text Add to dashboard Cite

show abstract

“…Box filters have shown to be highly discriminative for detecting smiles in natural image settings [62]. For the recognition of facial action units, the success is mixed: In a preliminary comparison [63] between box filters and Gabor filters on a subset of the Cohn-Kanade dataset, it was found that box filters yielded accuracies that were equally good as for the 2-D Gabors on just a few action units. On other action units, accuracy was significantly less.…”

Section: Box Filtersmentioning

confidence: 99%

Automatic Facial Expression Recognition

Whitehill¹,

Bartlett²,

Movellan³

2013

Social Emotions in Nature and Artifact

View full text Add to dashboard Cite

“…Although the classi cation rates for Gabor wavelets came out to be the most successful, Whitehill and Omlin [11] showed that extraction of Gabor wavelet coe cients is 300 times more costly than Haar. However, the down-sampled images or features might be used in real-time applications where speed is of importance.…”

Section: Binary Classi Cationmentioning

confidence: 99%

Upper Facial Action Unit Recognition

Zor

Windeatt

2009

Advances in Biometrics

View full text Add to dashboard Cite

Abstract. This paper concentrates on the comparisons of systems that are used for the recognition of expressions generated by six upper face action units (AU s) by using Facial Action Coding System (FACS ). Haar wavelet, Haar-Like and Gabor wavelet coe cients are compared, using Adaboost for feature selection. The binary classi cation results by using Support Vector Machines (SVM ) for the upper face AU s have been observed to be better than the current results in the literature, for example 96.5% for AU2 and 97.6% for AU5. In multi-class classi cation case, the Error Correcting Output Coding (ECOC ) has been applied. Although for a large number of classes, the results are not as accurate as the binary case, ECOC has the advantage of solving all problems simultaneously;and for large numbers of training samples and small number of classes, error rates are improved.

show abstract

Haar Features for FACS AU Recognition

Cited by 127 publications

References 7 publications

Active AU Based Patch Weighting for Facial Expression Recognition

Active AU Based Patch Weighting for Facial Expression Recognition

Automatic Facial Expression Recognition

Upper Facial Action Unit Recognition

Contact Info

Product

Resources

About