Title: MG-Gen: Single Image to Motion Graphics Generation with Layer Decomposition Show Chinese title

Title: MG-Gen：基于分层分解的单图像到运动图形生成

Abstract: General image-to-video generation methods often produce suboptimal animations that do not meet the requirements of animated graphics, as they lack active text motion and exhibit object distortion. Also, code-based animation generation methods typically require layer-structured vector data which are often not readily available for motion graphic generation. To address these challenges, we propose a novel framework named MG-Gen that reconstructs data in vector format from a single raster image to extend the capabilities of code-based methods to enable motion graphics generation from a raster image in the framework of general image-to-video generation. MG-Gen first decomposes the input image into layer-wise elements, reconstructs them as HTML format data and then generates executable JavaScript code for the reconstructed HTML data. We experimentally confirm that MG-Gen generates motion graphics while preserving text readability and input consistency. These successful results indicate that combining layer decomposition and animation code generation is an effective strategy for motion graphics generation. Show Chinese abstract

Abstract: 通用图像到视频生成方法往往会产生次优的动画，这些动画不符合动画图形的需求，因为它们缺乏主动的文本运动且存在物体失真现象。 此外，基于代码的动画生成方法通常需要分层结构的矢量数据，而这种数据在运动图形生成中往往不易获得。 为了解决这些挑战，我们提出了一种名为MG-Gen的新框架，该框架从单个光栅图像中重构矢量格式的数据，以扩展基于代码的方法的能力，从而能够在通用图像到视频生成框架中从光栅图像生成运动图形。 MG-Gen首先将输入图像分解为逐层元素，将其重构为HTML格式数据，然后为重构后的HTML数据生成可执行的JavaScript代码。 我们通过实验验证了MG-Gen能够生成运动图形的同时保留文本的可读性和输入的一致性。 这些成功的成果表明，结合分层分解和动画代码生成是一种有效的运动图形生成策略。