Local directional-structural pattern for person-independent facial expression recognition

Abstract

Existing popular descriptors for facial expression recognition often suffer from inconsistent feature description,experiencing poor accuracies. We present a new local descriptor, local directional-structural pattern (LDSP), in thiswork to address this issue. Unlike the existing local descriptors using only the texture or edge information to representthe local structure of a pixel, the proposed LDSP utilizes the positional relationship of the top edge responses of thetarget pixel to extract more detailed structural information of the local texture. We further exploit such information tocharacterize expression-affiliated crucial textures while discarding the random noisy patterns. Moreover, we introducea globally adaptive thresholding strategy to exclude futile flat patterns. Hence, LDSP offers a stable description offacial expressions with the explicit representation of the expression-affiliated features along with the exclusion of randomfutile textures. We visualize the efficacy of the proposed method in three folds. First, the LDSP descriptor possessesa moderate code-length owing to the exclusion of the futile patterns, yielding less computation time than other edgedescriptors. Second, for person-independent expression recognition in benchmark datasets, LDSP demonstrates higheraccuracy than existing descriptors and other state-of-the-art methods. Third, LDSP shows better performance thanother descriptors against noise and low resolution, exhibiting its robustness under such uneven conditions.