百度搜索
                      百度搜索
                      • 学院主页
                      • 信箱
                      • 捐赠
                      • 常用资源
                      • 旧版网站
                      • English

                      学术活动

                      首页 - 学术活动 - 正文

                      【葡京国际网站app40周年院庆系列学术活动第59期】复旦大学葡京国际网站app营销系副教授肖莉:智能视频广告推荐系统 An Intelligent Video Ad Recommendation (iVAR) System

                      • 发布日期:2019-11-22
                      • 点击数:

                        

                      喻园管理论坛2019年第139期(总第590期)


                      演讲主题:智能视频广告推荐系统 An Intelligent Video Ad Recommendation (iVAR) System

                      主 讲 人:肖  莉,复旦大学葡京国际网站app营销系副教授

                      主 持 人:阎  俊,葡京国际网站app工商管理系副教授

                      活动时间:2019年12月6日(周五) 15:00-16:30

                      活动地点:葡京国际网站app119教室

                      主讲人简介:

                      肖莉现任复旦大学葡京国际网站app营销系副教授。她毕业于宾州州立大学Smeal商学院,获得营销学博士学位。她在Marketing Science上发表过二篇学术论文,承担过一项国家自然科学基金青年项目。她的研究兴趣集中在运用图像视频数据和计算机视觉技术解决营销工作中的问题,以及多媒体数据在广告、产品创新及方法设计上的运用。

                      活动简介:

                      In this paper, we propose an intelligent video ad recommendation (iVAR) system to provide a customized delivery of video ads. The iVAR system builds on a visual object based preference model and utilizes facial expression-eye gaze (abbreviated as EG) stream data that track the facial expression and eye gaze of a viewer while watching an ad by recording the viewing experience and extracting and analyzing at the frame level. By tracking viewers’ facial responses in response to various visual objects in real time, the system can infer a viewer’s preferences towards different ads, search the ad database, and select and subsequently display a new ad that is most likely to achieve positive attitudinal and behavioral consequences. We demonstrate the feasibility and performance of the proposed system with two empirical studies. The results show that by tracking viewers’ facial responses to only one ad or even part of one ad, our proposed system is able to make reasonably accurate inferences of the viewer’s preferences towards video ads, with or without using other viewers’ information, and make recommendations that help achieve favorable attitudinal and behavioral responses.

                      学院要闻

                      • 1
                      • 2
                      • 3
                      • 4
                      • 5