摘要
随着计算机技术和人工智能研究的不断进步,表情识别(Facial Expression Recognition, FER)是人工心理和情感研究的核心问题,也是人机交互技术的重要方向。本文深入探讨了表情识别系统的设计,特别是对特征提取相关算法进行了改进,目的是提高表情识别的精度,并最终实现表情识别系统的目标。本文使用一种基于改进的AdaBoost算法和Haar特征的方法,用于提高人脸检测的准确率性,并且提高速度。对比改进前后算法训练时长和准确率,梳理新旧算法在人脸检测的实验结果,证实了新算法保证准确率的基础上,缩短训练时间,具有一定的优越性。
With the continuous progress of computer technology and artificial intelligence research, Facial Expression Recognition (FER) is the core problem of artificial psychology and emotion research, and it is also an important direction of human-computer interaction technology. This paper discusses the design of the expression recognition system in depth, especially the improvement of the feature extraction-related algorithm, in order to improve the accuracy of expression recognition and ultimately achieve the goal of the expression recognition system. This paper uses a method based on the improved AdaBoost algorithm and Haar features to improve the accuracy and speed of face detection. Comparing the training time and accuracy rate of the algorithm before and after improvement, and combing the experimental results of the new and old algorithms in face detection, it is confirmed that the new algorithm has certain advantages in shortening the training time on the basis of ensuring accuracy.
出处
《图像与信号处理》
2023年第3期260-269,共10页
Journal of Image and Signal Processing
