随着计算机技术与交互方式的发展进步，越来越多服务系统采用智能媒体计算技术来方便用户获取信息。而传统智能媒体计算方法在用户部分信息通道缺失的前提下无法提供较好的服务，因此惠及老年人、残疾人等特殊人群的无障碍智能媒体计算技术研究成为国际学术研究的热点之一。申请人拟针对面向视障用户的多通道媒体智能计算技术中的语义鸿沟、信息过载、同步交互三大科学问题开展研究，以无障碍智能媒体计算服务的理论和关键技术为支撑，借鉴群体智能、自主智能等新一代人工智能的优势，研究基于群体智能与大数据的智能媒体计算方法，广泛利用互联网用户的力量跨越单靠计算机难以彻底解决的语义鸿沟问题，探索无障碍智能媒体特定场景表达与理解、无障碍智能媒体群体智能协同计算、无障碍智能媒体自主学习反馈交互等研究工作，并通过典型应用完善无障碍智能媒体计算服务体系，实现更高质量、更高性能的无障碍智能媒体计算。

With the development of computer technology and interaction methods, more and more service systems use intelligent media computing technology to facilitate users to obtain information. However, the traditional intelligent media computing method can not provide better services under the premise of the lack of user information channels. Therefore, the research on accessible intelligent media computing technology that benefits the elderly and disabled people has become one of the hotspots of international academic research. The applicant intends to conduct research on the three major scientific issues of semantic gap, information overload and synchronous interaction in accessible intelligent media computing, aiming at improving the ability of accessible intelligent media computing services, and the theory and key technologies of accessible intelligent media computing services. To support and learn from the advantages of a new generation of artificial intelligence such as crowdsourcing intelligence and autonomous intelligence, research on intelligent media computing methods based on group intelligence and big data, and widely use the power of Internet users to cross the semantic gap problem that is difficult to solve completely by computers alone. Obstacle intelligent media specific scene expression and understanding, barrier-free intelligent media group intelligent collaborative computing, accessibility intelligent media independent learning feedback interaction and other research work, and through the typical application to improve the accessible intelligent media computing service system to achieve higher quality, higher Performance of accessible smart media computing.