基于CNN集成学习的人脸表情识别系统的设计

Design of Facial Expression Recognition SystemBased on CNN Integrated Learning

随着计算机计算资源的提升以及深度学习理论的不断丰富,自动的人脸表情识别技术已经得到了进一步的发展。但由于表情存在复杂性以及微妙性,实现实时的人脸表情识别仍是一大难题。文章设计了一种基于CNN集成学习的人脸表情识别系统,该系统在FER2013数据集上表情的识别准确率达到70.84%,能够实现实时的、高精度的表情识别。

With the improvement of computer computing resources and the continuous enrichment of deep learning theory,automatic facial expression recognition technology has been further developed.But because of the complexity and subtlety of facial expressions,it is still a big problem to realize real-time facial expression recognition.Therefore,we design a facial expression recognition system based on CNN integrated learning in this paper,which can realize real-time and high-performance facial expression recognition with an accuracy rate of 70.84%on FER2013 data set.