iQIYI and ACM MM Launch 2019 Celebrity Video Identification Challenge

Last year’s inaugural competition was held using the largest celebrity video dataset in the world at the time, which enabled iQIYI to upgrade its technology and achieve a breakthrough in video-based recognition technology. This year, iQIYI will provide an upgraded data set (iQIYI-VID-2019) of 200,000 video segments containing 10,000 human figures, 5,000 of which come from the iQIYI celebrity database and iQIYI-VID-2018. The other 5,000 figures are extracted from iQIYI’s short video App iQIYI Jiangbing, representing 200 hours of new content. Researchers will compete to develop person recognition algorithms to identify celebrities through multimodal features such as the face, voice, movement and clothing in the videos.

Marketing Technology News: Quantum Names Liz King Chief Revenue Officer

“We are delighted to organize this year’s challenge with the prestigious ACM MM and we look forward to promoting breakthroughs of multimodal person recognition research to apply academic research of computer vision (CV) and AI to industrial use,” said Liu Wenfeng, Chief Technology Officer at iQIYI. “We are confident the competition will yield AI breakthroughs with practical application to further optimize iQIYI’s entertainment ecosystem and monetization model while improving user experience.”

iQIYI is committed to enhancing the application of AI in the entertainment industry through research cooperation. This commitment is reflected in the company’s recently released 2018 financial results, in which R&D expenses were US$290.1 million, representing a 57% increase from 2017. In the future, AI capabilities will remain a key focus of iQIYI as the company continues its journey to become a technology-based entertainment leader.

Marketing Technology News: Qualtrics Partner Network Grows to over 100 Global Members and Expands into Five Programs in Its First Year of Revolutionizing Customer, Employee, Brand and Product Experiences for Organizations Worldwide

ACM MM is a premier international conference on multimedia research including images, text, video, music, sensor data, spoken audio and social network. Since 1993, ACM MM has been bringing together leading researchers and practitioners from academia and industry, such as Carnegie Mellon University, Columbia University, MIT Media Lab, Google, Microsoft and IBM, to present innovative research results and discuss recent advancements.