摘要
视觉场景理解是机器视觉与自然语言处理相融合的热点问题,实现场景理解将有助于盲人辅助系统、智能机器人交互、以及无人驾驶等领域研究的发展。从视觉场景理解的基本概念出发,介绍了视觉场景理解技术的相关理论及最新研究成果,并对基于搜索的模型、基于模板匹配的模型和基于编码-解码框架的模型、以及基于注意力模型的模型进行详细描述与分析。讨论视觉场景理解的最新技术动态,分析视觉场景理解的发展趋势并指出未来研究方向。
Image caption is a hot topic in the fusion of machine vision and Natural Language Processing . The realization of image caption will help the development of blind people auxiliary system, intelligent robot interaction, and unmanned driving. Some basic concepts of image caption are introduced, the search-based model, template matching-based model, coding-decoding framework-based model and attention mechanism-based model are described and analyzed in detail . The latest technology trends of visual image captioning, the development trend of visual image captioning and several future research directions are also discussed.
作者
王忠民
王星
李刚
张福涛
WANG Zhongmin;WANG Xing;LI Gang;ZHANG Futao(School of Computer Science and Technology, Xi'an University of Posts and Telecommunications,Xi'an 710121,China;Shaanxi Key Laboratory of Network Data Intelligent Processing, Xi'an University of Posts and Telecommunications, Xi'an 710121,China)
出处
《西安邮电大学学报》
2019年第1期1-15,共15页
Journal of Xi’an University of Posts and Telecommunications
基金
国家自然科学基金资助项目(61373116
61702414)
陕西省工业科技攻关资助项目(2016GY-092)
陕西省科技统筹创新工程计划资助项目(2016KTZDGY04-01)
陕西省教育厅专项科学研究计划资助项目(17JK0711)
