The rapid developments of artificial intelligence have significantly impacted daily life and content production modes. In the field of video generation, researchers are now exploring this emerging technique with innovative approaches, aiming to produce videos of higher quality, longer duration, and greater diversity. Currently, numerous video generation algorithms have been developed using different architecture designs. Unlike image generation, video generation requires maintaining consistency across both spatial and temporal dimensions while ensuring aesthetic quality and dynamic coherence, making it a more challenging task. In this survey, we provide a systematic review of existing video generation methods, tracing their evolution across different architectural paradigms. We further categorize recent models by their control conditions (e.g., text-to-video, image-to-video, multi-modal guidance) and summarize their unique theoretical foundations, architectural designs, and algorithmic innovations. In the meantime, we review the commonly used video datasets and analyze their applicability to different tasks. We also present evaluations of representative models to offer a more comprehensive perspective. Our goal is to provide a clear and concise overview of these algorithms, offering insights to support future breakthroughs in video generation.

We present a unified controllable video generation approach AnimateAnything that facilitates precise and consistent video manipulation across various conditions, including camera trajectories, text prompts, and user motion annotations. Specifically, we carefully design a multiscale control feature fusion network to construct a common motion representation for different conditions. It explicitly converts all control information into frame-by-frame optical flows. Then we incorporate the optical flows as motion priors to guide the final video generation. In addition, to reduce the flickering issues caused by large-scale motion, we propose a frequency-based stabilization module. It can enhance temporal coherence by ensuring the video's frequency domain consistency. Experiments demonstrate that our method outperforms the state-of-the-art approaches. For more details and videos, please refer to the anonymous webpage: https://yu-shaonian.github.io/Animate_Anything/.

Generating videos guided by camera trajectories poses significant challenges in achieving consistency and generalizability, particularly when active objects are present. Existing approaches often attempt to learn object motions separately, which may lead to confusion regarding the relative motion between the camera and the objects. To address this challenge, we propose a novel approach that integrates both camera trajectory and object semantics and converts them into the reference motion of the corresponding pixels. Utilizing a stable diffusion network, we effectively extract reference motion maps in relation to the specified camera trajectory. These maps, along with an extracted semantic object prior, are then fed into an image-to-video network to generate the desired video that can accurately follow the designated camera trajectory while maintaining the consistency of object. Extensive experiments verify that our model outperforms SOTA methods by a large margin. Project Page: https://yu-shaonian.github.io/MotionFlow/

Recent advancements in video generation models have enabled the creation of diverse and realistic videos, with promising applications in advertising and film production. However, as one of the essential tasks of video generation models, video concept transfer remains significantly challenging. Existing methods generally model video as an entirety, leading to limited flexibility and precision when solely editing specific regions or concepts. To mitigate this dilemma, we propose a novel architecture UniTransfer, which introduces both spatial and diffusion timestep decomposition in a progressive paradigm, achieving precise and controllable video concept transfer. Specifically, in terms of spatial decomposition, we decouple videos into three key components: the foreground subject, the background, and the motion flow. Building upon this decomposed formulation, we further introduce a dual-to-single-stream DiT-based architecture for supporting fine-grained control over different components in the videos. We also introduce a self-supervised pretraining strategy based on random masking to enhance the decomposed representation learning from large-scale unlabeled video data. Inspired by the Chain-of-Thought reasoning paradigm, we further revisit the denoising diffusion process and propose a Chain-of-Prompt (CoP) mechanism to achieve the timestep decomposition. We decompose the denoising process into three stages of different granularity and leverage large language models (LLMs) for stage-specific instructions to guide the generation progressively. We also curate an animal-centric video dataset called OpenAnimal to facilitate the advancement and benchmarking of research in video concept transfer. Extensive experiments demonstrate that our method achieves high-quality and controllable video concept transfer across diverse reference images and scenes, surpassing existing baselines in both visual fidelity and editability.