The 2024 World Artificial Intelligence Conference opened in Shanghai today, and Vimi, the first controllable character video generation model designed by Shangtang Technology for C-end users, was selected as the highest honor "Treasure of the Town" in the WAIC exhibition, becoming the most innovative exhibit of this conference. Vimi is based on the powerful capabilities of the Shangtang Ri Ri Xin model. It can generate character videos that match the target action with just one photo of any style, and supports multiple driving methods. It can be driven by various elements such as existing character videos, animations, sounds, and text. With the rapid development of large models and generative AI technology, making the characters in photos move is no longer a new thing. But currently, there are many pain points in the practical application of related products on the market. Some character movements and facial expressions cannot be precisely controlled, and can only be frequently tried with various Prompt card draws; Some effects have low stability, and the appearance and background effects of the characters are unpredictable; Some videos have limited duration and can only generate 3-4 second videos, which limits their creativity and cannot meet the needs of video creators. In order to solve the practical problems in these applications, the large model Vimi for controllable character video generation in Shangtang has emerged. Unlike image expression control techniques that can only control head expression movements, Vimi not only achieves precise control of character expressions, but also controls the natural body movements of characters in photos within the body area, and automatically generates hair, clothing, and background changes that match the characters. At the same time, the changes in light and shadow can also be reasonably generated, making the character's movements and visual effects smooth and natural, and the picture harmonious and beautiful. More importantly, Vimi has strong stability and can stably generate single shot character videos for up to 1 minute. The visual effect will not deteriorate or distort over time, truly meeting the needs of long-term stable video generation for entertainment and interaction. Easy to become an internet celebrity and meet the entertainment and creative needs of users, Vimi will be fully open to C-end users, especially to meet the entertainment and creative needs of female users. Users only need to upload high-definition character images from different angles to automatically generate digital avatars and different styles of photo videos. The video characters generated by Vimi are no longer just rigid facial movements, but are combined with gestures, limbs, hair, etc. to form more complete and unified character movements, allowing creators to edit and recreate based on the generated video materials. For users who love taking selfies, Vimi can support various entertainment interactive scenes such as chatting, singing, dancing, etc; For users who are passionate about emoticons, Vimi can generate various interesting character emoticons through a single image, with diverse gameplay and creative freedom. In addition, Vimi also offers various generative styles such as beautiful photo style and fantasy style, allowing users to experience immersive visual effects of blockbuster textures as if traveling through different dimensions. At present, 80% of videos in the industry are mainly about characters, and with the explosion of short videos and live streaming platforms, the demand for videos with characters as the main body is growing exponentially. However, the lack of AIGC products for controllable character video generation on the market has led to a need to improve the efficiency of content creators. The emergence of Vimi precisely fills this demand gap, providing a simple and useful creative tool for video creators on the C-end, which can be widely applied to the creation of short videos about characters. (Lai Xin She)