Author(s)

Chenhao Si

Corresponding Author

Chenhao Si

ABSTRACT

The existing facial expression recognition technology is basically limited to the traditional machine learning algorithm. In the case of light intensity, occlusion and attitude transformation, the traditional machine learning algorithm has poor robustness and is difficult to apply to real life. With the development of hardware conditions such as computer GPU and the arrival of the era of big data, in-depth learning has attracted much attention in the field of computer vision. The traditional machine learning algorithms and their advantages and disadvantages are introduced in terms of processing, feature extraction and feature classification. The deep learning algorithms are introduced from DBN, CNN and other mainstream algorithms, development directions, common development frameworks. Finally, the development problems and trends of traditional machine learning and deep learning in facial expression recognition are summarized and prospected, as well as the future research directions.

KEYWORDS

Facial expression recognition, Deep learning, CNN, Machine learning, Computer vision, Image preprocessing, Feature extraction, Feature classification