Thursday 24 Nov 2016: Processing and Analysis of Multimedia Big Data
Prof. Ruimin Hu - Wuhan University, China
Harrison 170 14:30-15:30
Multimedia, as one of the most important and valuable sources for insights and information, is increasingly being considered as big data. Multimedia big data inherits the characteristics of both multimedia and big data, leading to tremendous challenges in data processing and analysis. For example, the explosion of data amount leads to high pressure on data storage and transmission, calling for a high efficient data representation of multimedia data. Moreover, accuracy of conventional data analysis drops dramatically with the increment of complexity and variety of data, requiring new techniques for intelligent analysis for multimedia big data. Most importantly, a data mining technology is in highly demand to discover valuable information embedded in the large-heterogeneous data. Based on the above consideration on multimedia big data, this presentation will mainly cover (a) high efficient data compression of video big data; (b) accurate representation of 3D audio data; (c) high accurate people re-identification; and (d) customer loyalty assessment.