The decadal perspective of facial emotion processing and Recognition.

What is it about?

Facial expression recognition (FER) is playing a crucial role in distinct psychological disorders, human–machine interaction, and a multitude of multimedia applications. The transformation of FER from lab to wild conditions and significant advancement in deep learning has led to the implementation of automatic FER. In this article, we provide a review of FER that includes Ekman’s six basic emotions, the significance of FER with datasets, and deep learning algorithms. The article classified the fundamental procedure of FER into distinct for clear understanding. The significance of each procedure in FER including face detection& tracking, extracting facial features of dynamic & static images, and facial expression classification is addressed with algorithms in this article. The existing state of art deep neural networks including convolution neural network (CNN), deep belief network (DBN), the deep auto encoder (DAE), and recurrent neural network (RNN) for FER are also presented in this article. Finally, the article provides challenges and recommendations namely deficiency in datasets, biasness and inconsistency in data set, integration of robust models, multimodal for effective recognition. FER with technology for sustainable health, edge computing powered devices for FER implementation, adoption of FER-based human interaction robots, and customized portable device for FER.

Featured Image

Why is it important?

Facial expression is one of the fundamental, influential, natural, and universal gestures for human beings to express their emotional state and impulsion. Facial emotion recognition (FER) has gained significant interest in the field of human–computer interaction, medical treatment, sociable robots, and driver warning systems.

Perspectives