Comparison on machine learning based algorithms for face expression recognition
Jiayi Bai, Jiawei He, Fengyuan Liu · Applied and Computational Engineering · 2023
With the development of machine learning methods, most of tasks has benefited from it. Numerous works have been done to improve the accuracy and robustness of these methods. In this paper, the authors aim to exploit the effectiveness of machine learning algorithms in face expression recognition (FER) task. Three methods, including support vector machine, random forest and convolutional neural network are introduced to test the performance in FER task. Three methods are tested in the public dataset, FER2013. The conclusion is drawn as follow. All these methods can handle the image recognition, which is a 7-class classification task. Furthermore, nerual network achieves the best performance, which achieves 65% accuracy. The paper exploit the potential of machine learning methods and give a brief attempt on applying the popular algorithms into real-world image recognition tasks.