BEIJING, Feb. 16, 2022 /PRNewswire/ — Moviebook, a leading AI-powered infrastructure and service provider of intelligent video production, recently released its AI-powered event content generation solution — Moviebook SAiDT — to help event organizers build one-stop platforms for intelligent video content, enabling the production and operation of visually enhanced content while significantly reducing content generation costs, and improving organizers’ ability to provide various types of content simultaneously.
Developed by integrating Generative AI and digital twin technology, Moviebook SAiDT features the Company’s self-developed AI tool stack including an estimation framework used to capture motion from 2D video (MCVS) and an infrastructure platform for immersive media content creation (MAGC). The SAiDT solution supports track simulation, stadium simulation, and simulation of indoor navigation in the stadiums, which combined with event content can generate new enhanced competition video.
Moviebook SAiDT was successfully applied to test events at the winter games in Beijing and contributed to the goal of efficiently providing quality sports competition videos for the audience. Meanwhile, a system for automatic recognition of sports event video, built on Moviebook’s AI visual recognition technology, can be leveraged to identify live streaming signals of events in real time, automatically compress large amounts of video content into several minutes of sports highlights and publish these highlights in a targeted manner. A model characterized by efficient, accurate content production at scale greatly reduces manpower and time costs and opens up more possibilities for further unleashing the value of sports video.
In addition, Moviebook SAiDT supports the import of video streams from existing devices and the one click upgrade of 2D video to AI-enhanced video, taking into account both the cost-effectiveness and operational efficiency of a large amount of AI video content. After standard video streams are analyzed and key frames extracted, high-tech annotators will be used to label data more accurately and to automatically generate new key data. Subsequently, new AI-enhanced video will be created and transmitted to the media platform, synchronizing content generation and commercialization.
In fact, automatic generation and operation of event content is one of the three top priorities of a high-tech winter games. With its emphasis on the integration of practicality and technology, Moviebook SAiDT divides the automatic generation and operation flow into four parts: front-end video import, automatic recognition and motion capturing estimation, automatic generation, and self-adaptation to operational scenarios. Buttressed by its advanced AI content generation technology, Moviebook SAiDT realizes the automation of event video generation and operation and upgrades traditional video recording and editing processes with a high threshold and massive investment to AI-powered processes. This solution has been adopted by major cultural and sports content ecosystems.
Moviebook SAiDT, as a part of the AI-based content generation infrastructure MAGC (Moviebook Automatic Generated Content), is customized for sports events. Building on mature AI algorithms such as digital bionic interaction and 3D modeling and reconstruction, and stable AI engineering ability, the solution can help quickly generate visually enhanced interactive video content, customize athlete training models and sports highlights, and automatically edit video into slick short video based on an understanding of the content.
According to the head of an event operation platform which has adopted the SAiDT solution, real major changes will come when a technology is not only used for entertainment or research but also widely applied in practice, because similar technologies may become standards of various sports events. Data shows that after SAiDT is employed to analyze event images, automatically complete pixel by pixel annotation, and identify key features for 3D pose estimation, simplified 3D rendering of models can be completed, and AI-based video content can be generated. The adoption of Moviebook SAiDT can considerably lower labor costs by automating the whole process spanning from capturing key shots to content generation and content delivery.
A global consensus has also been reached on mainstreaming AI-based event broadcasting. Statistics from Zhongshang Institute for Industry Studies show that China’s sports event operation market in 2021 reached US$57.39 billion (RMB 363.67 billion). AI-based automatic content production has become an essential prerequisite for the continued growth of this promising market. Thanks to advances in computing technology, 2D images can be converted into immersive models more effectively, and optimal content can be provided to consumers most effectively using AI. In the future, athletes may rely more on massive AI-processed data to improve their competition performance; event operation platforms can also achieve improved development through AI-based automatic content generation and broadcasting.
Moviebook, a leading AI-powered infrastructure and service provider of intelligent video production, aims at empowering monetization capacity for clients from new retail, media, education, and culture through intelligent vision. Building on technological advantages in the fields of computer vision, computer graphics, among others, Moviebook significantly improves the production efficiency of visual content and innovates ways of presenting and interacting with such content. By applying AI technologies to videos, Moviebook is pioneering a wide variety of application scenarios for intelligent video production technologies.
For more information, please visit: http://www.moviebook.cn/about/index.