Recently, Kling introduced a brand new "Custom Models" feature, a significant update that will thoroughly address the technical challenges of character consistency in AI video generation. Based on feedback from early testing users on the platform, this feature allows creators to upload 10 to 30 different video clips for model training, with each clip requiring at least 10 seconds. Unlike traditional models that rely solely on image training, KLING innovatively utilizes video materials as the training basis and even supports using AI-generated videos created by Kling for training.

China's first fully self-developed video model, Vidu, created by Shengshu Technology and Tsinghua University, introduces the 'Subject Reference' feature to ensure character consistency in AI-generated videos. Users can upload images, and Vidu maintains consistency across various scenes. This free feature marks a new era in AI video creation, aiming for precise multi-subject interaction and style uniformity. Director Shi Yuxiang praised its enhanc....

The research team at Alibaba Group has released a new framework called UniPortrait, which focuses on personalized processing of portrait images, including single-character consistency, multi-character consistency, and style reference. UniPortrait is specifically designed with two modules: ID embedding and ID routing, which maintain a high level of realism in facial features, support a broad range of facial editing capabilities, and even allow users to generate images using free-form text descriptions without a fixed layout. Through a two-phase training scheme, UniPo