基于深度学习的多模态动作识别方法探究

Research on Multi-modal Action Recognition Method Based on Deep Learning

在人工智能与大数据技术不断发展的背景下,信息的来源与呈现形式都产生较大程度的变化,多元化的媒介形态与信息为大众带来了多重的视听体验,多模态机器学习应运而生。笔者主要针对深度学习理论以及多模态动作识别理论进行阐述与解读,并对深度学习与多模态动作识别之间的关联性进行分析,从而在此基础上探究深度学习的多模态动作识别方法,以此推进我国媒介技术发展的研究进程。

Under the background of the continuous development of artificial intelligence and big data technology,the sources and presentation forms of information have undergone major changes.Diversified media forms and information have brought multiple audiovisual experiences to the public.State-of-the-art machine learning came into being.This article mainly elaborates and interprets deep learning theory and multi-modal action recognition theory,and analyzes the correlation between deep learning and multi-modal action recognition,so as to explore the multi-modality of deep learning on this basis Action recognition method to promote the research process of the development of media technology in our country.