ICCV Daily 2021 - Tuesday

Currently, the team are focused on achieving the best performance of the visual appearance . In terms of future work, they would like to improve the efficiency of the model to make it faster for real-time or near real-time processing. They would also like to consider application-specific settings, so if there is a very heavy compression or a light compression, the training can be targeted to the specific settings for the best output results. “ I think the paper opens the door to new areas for video super resolution, ” Yinxiao tells us. “ I would encourage researchers to focus on more realistic problems. For example, videos in the wild. They can make such videos as testing videos for their new video super-resolution work or any other enhancement model so that this can be easily applied to practical problems. ” Outside of this paper, Yinxiao works at Google Research , where his team are developing fundamental state-of-the-art imaging technology which is widely applicable to all relevant Google products. “ High-quality multimedia content is very important for advertising, ” he says. “ My work is mainly targeted at enhancing that multimedia content. It provides a better user experience for customers. The nicer it is, the happier the customer will feel. ” So, when we see advertising on YouTube , it will be faster and flow better? There are more than 2 billion YouTube users that say thank you, Yinxiao! He laughs – “ Yes, I hope so! ” To learn more about Yinxiao’s work [Paper ID 10703], you can view his pre-recorded oral presentation and PDF poster now and join his live Q&A sessions today [2A] at 11:00-12:00 EDT and Thursday [2B] at 18:00-19:00 EDT. 10 DAILY ICCV Tuesday Oral Presentation