深度学习下的行人再识别问题研究

Research of Person Re-identification Based on Deep Learning Method

行人再识别是当前大型智能监控系统中的关键技术,主要用于在监控视频中查询与匹配特定的目标行人,近年来随着计算机视觉、深度学习技术的发展,行人再识别技术从传统方法逐渐过渡到深度学习方法成为主流。笔者从基于深度学习的行人再识别研究现状出发,从特征学习和测度学习两个方面就提高行人再识别算法的准确性的关键技术进行研究。

Person Re-identification is a key technology in current large-scale intelligent monitoring systems.It is mainly used to query and match specific target pedestrians in surveillance videos.In recent years,with the development of computer vision and deep-learning technologies,pedestrian re-identification technology has gradually evolved from traditional methods to deep learning methods.The author starts from the research status of person re-identification based on deep learning,and studies the key technologies to improve the accuracy of person re-identification algorithm from the aspects of feature learning and distance metric learning.