https://www.solosoft.dev/tags/nankai/Recent content in Nankai on SoloSoftHugoen-usFri, 01 May 2026 00:00:00 +0000https://www.solosoft.dev/post/storydiffusion-image-video-2026/Fri, 01 May 2026 00:00:00 +0000https://www.solosoft.dev/post/storydiffusion-image-video-2026/<p><strong>StoryDiffusion</strong> is a research project from Nankai University and ByteDance that tackles one of the hardest problems in generative AI: maintaining visual consistency across long sequences of images and videos. Accepted as a major research contribution, it introduces a novel <strong>consistent self-attention (CSA)</strong> mechanism that enables diffusion models to generate coherent comic strips, animations, and videos &ndash; all without finetuning or per-sequence training.</p> <p>The core challenge StoryDiffusion addresses is simple to state but extremely difficult to solve: how do you generate a sequence of images where the same character looks consistently the same in every frame? Previous diffusion models could produce stunning single images, but when asked to generate a multi-panel comic or a video clip, characters would subtly change appearance between frames &ndash; a different nose shape, a changed outfit, a shifted background style.</p>