应信息与系统科学研究所邀请，英特尔中国研究院主任研究员李建国博士将在我院作学术报告：

        时间：2015年9月21日下午3:00
        地点：理科楼122
        报告题目：Context Aware Deep Attributes for Visual Recognition

        报告摘要: Recently, many researches employ mid-layer neural codes of convolutional neural network models as feature representations for visual recognition tasks. Although empirical success being observed in practice, this type of representations have limited capacity in exploring the context information within images, resulting limited performance gain due to the well-known semantic gap issue. To address such a challenge, we propose a "context-aware deep attribute" framework that is able to synergistically combine effective neural coding and semantic region information. The proposed deep attributes have been evaluated on various vision recognition tasks. Experimental results corroborate the effectiveness of the proposed technique by showing a significant performance gain over state-of-the-art methods.

        个人简介：李建国博士是英特尔中国研究院主任研究员。他2001年毕业于华中科技大学，2006年在清华大学自动化系获得博士学位。研究兴趣包括计算机视觉，机器学习和并行计算，以及它们的实际应用。他的诸多研究成果转换为英特尔公司产品特征，包括CPU/GPU硬件特性，英特尔感知计算SDK等等。他获得30多项美国专利授权，发表了30多篇学术文章，其中包括10余篇顶级会议如CVPR, ICCV, ICML, IJCAI, ACM MM的文章。