一种运镜方式可控的视频生成方法、装置及电子设备
By constructing a joint training mechanism for the camera movement information encoder and the video generation model, the problems of inaccurate camera movement control and insufficient generation quality in existing technologies are solved, and high-quality video generation with controllable camera movement is achieved.
CN121099153BActive Publication Date: 2026-07-17CHENGDU SOBEY DIGITAL TECH CO LTD
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CHENGDU SOBEY DIGITAL TECH CO LTD
- Filing Date
- 2025-08-14
- Publication Date
- 2026-07-17
AI Technical Summary
Technical Problem
Existing video generation technologies lack precise camera movement control and high-quality generation methods, making it difficult to meet diverse video generation needs.
Method used
A camera movement information encoder and a video generation model are jointly trained. Camera movement videos are generated by generating time-series camera pose parameters, images, and text descriptions. Deep neural networks are used to achieve controllability of camera movement methods.
Benefits of technology
It achieves precise camera movement control and high-quality video generation, enhancing the flexibility and practical value of video generation.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN121099153B_ABST
Abstract
本申请的实施例提供了一种运镜方式可控的视频生成方法、装置及电子设备，涉及视频生成和深度学习技术领域，所述方法首先根据选择的运镜方式生成时序的相机位姿参数，然后将相机位姿参数进行预处理并输入运镜信息编码器得到运镜信息编码特征，最后将运镜信息编码特征、图片以及文本描述输入视频生成模型中生成对应的运镜视频。该方法构建了一个运镜信息编码器，并且和视频生成模型进行联合训练，使得原始的视频生成模型可以接受运镜信息，从而实现了运镜方式可控的视频生成，充分利用了视频生成模型的创造性并提供了精确的运镜控制。
Need to check novelty before this filing date? Find Prior Art