Image-to-Video¶
Video
Open Source¶
-
Helios Discovered by AI
2026
Helios is a 14B parameter autoregressive video generation model designed for real-time, long-duration video synthesis. It runs at 19.5 FPS on a single H100 GPU, supports both text-to-video and image-to-video tasks, and maintains high temporal consistency over minute-scale generations without drifting.
autoregressiveimage-to-videolong-videoreal-time -
Dynamic-I2V Discovered by AI
2025
An innovative image-to-video generation framework that integrates Multimodal Large Language Models (MLLMs) to jointly encode visual and textual conditions, significantly enhancing the understanding of complex motion and object-action relationships.
diffusion-transformerimage-to-videomllmmotion-understanding -
HunyuanCustom Discovered by AI
2025
A multimodal-driven customized video generation framework built on HunyuanVideo that supports image, audio, video, and text conditions. It focuses on maintaining subject identity consistency and enabling flexible control over motion and content.
customizationimage-to-videomultimodalvideo-generation -
HunyuanVideo-I2V Discovered by AI
2025
An open-source, customizable image-to-video generation framework built on the HunyuanVideo foundation model, supporting high-quality video synthesis up to 720p and 129 frames. It supports custom LoRA training for specialized visual effects and integrates a pre-trained MLLM for enhanced semantic understanding.
diffusionimage-to-videovideo-generation -
InfiniteTalk Discovered by AI
2025
InfiniteTalk is an open-source, unlimited-length talking video generation model that supports both audio-driven video-to-video and image-to-video generation. It overcomes the duration limitations of traditional talking head models while maintaining high lip-sync accuracy and identity stability.
audio-drivenimage-to-videolong-videotalking-head -
Pusa V1.0 Discovered by AI
2025
A versatile video diffusion framework that introduces Vectorized Timestep Adaptation (VTA) to enable fine-grained temporal control and efficient image-to-video generation, surpassing larger models with minimal training costs.
diffusionimage-to-videotemporal-controlvideo-generation -
Stable Virtual Camera: Generative View Synthesis with Diffusion Models Discovered by AI
2025
A 1.3-billion parameter generalist diffusion model (Seva) for Novel View Synthesis that generates 3D-consistent novel views of a scene from a single or multiple input images, allowing precise camera trajectory control.
camera-controldiffusionimage-to-videonovel-view-synthesisProject · GitHub · arXiv · Hugging Face
-
Wan2.1 Discovered by AI
2025
An open-source suite of video foundation models developed by Alibaba, featuring advanced text-to-video and image-to-video generation capabilities with high temporal consistency and motion quality.
diffusion-transformerimage-to-videovideo-generation -
CogVideoX Discovered by AI
2024
A large-scale open-source text-and-image-to-video diffusion model featuring a modality-adaptive expert transformer architecture for generating highly coherent and semantically aligned videos.
diffusionimage-to-videotransformer -
ConsisID Discovered by AI
2024
A tuning-free identity-preserving video generation framework that leverages frequency decomposition and DiT-based control to animate portrait images with highly consistent facial features.
ditidentity-preservingimage-to-videoportrait-animation -
ConsistI2V Discovered by AI
2024
A diffusion-based image-to-video generation framework designed to enhance visual consistency across frames. It introduces spatial-temporal attention mechanisms to preserve the identity of subjects, background details, and artistic style from the first frame.
consistencydiffusionimage-to-videovideo-generation -
DragAnything Discovered by AI
2024
A controllable video generation framework that utilizes entity representations to achieve precise trajectory-based motion control for any object in an image.
image-to-videomotion-controltrajectory-control -
EasyAnimate Discovered by AI
2024
An end-to-end open framework based on Diffusion Transformers for high-resolution, long image-to-video and text-to-video generation using hybrid windows attention and reward backpropagation.
diffusion-transformerimage-to-videolong-videoopen-source -
Follow-Your-Click Discovered by AI
2024
An image-to-video generation framework that enables precise regional animation and motion control of specific objects in an image based on user clicks and short text prompts.
image-to-videomotion-controlregional-animation -
LTX-Video Discovered by AI
2024
A transformer-based latent diffusion model that integrates a Video-VAE and a denoising transformer to generate high-quality videos in real-time. It supports both text-to-video and image-to-video generation at high resolutions and frame rates.
ditimage-to-videoreal-timevideo-generation -
MOFA-Video Discovered by AI
2024
MOFA-Video is a controllable image animation framework that adapts motions from different domains to a frozen image-to-video diffusion model. By employing sparse-to-dense motion generation and flow-based motion adaptation, it enables precise control over image animation using trajectories, keypoint sequences, or reference videos.
controllable-generationimage-animationmotion-transfer -
Motion-I2V Discovered by AI
2024
A consistent and controllable image-to-video generation framework that factorizes the process into explicit motion field prediction and motion-augmented temporal attention.
image-to-videomotion-modelingtemporal-attention -
Open-Sora Discovered by AI
2024
An open-source initiative by Colossal-AI to democratize efficient, high-quality video production, supporting text-to-video, image-to-video, and video-to-video generation with flexible aspect ratios and durations.
diffusionimage-to-videoopen-sourcevideo-generation -
Open-Sora Plan Discovered by AI
2024
An open-source large video generation framework featuring a Wavelet-Flow Variational Autoencoder (WFVAE) and supporting high-resolution, long-duration text-to-video and image-to-video synthesis.
diffusionimage-to-videovaevideo-generation -
AnimateDiff Discovered by AI
2023
A practical framework to animate personalized text-to-image diffusion models without specific tuning, widely used for generating stylized animations and video clips from static images.
animationimage-to-videostable-diffusion -
DynamiCrafter Discovered by AI
2023
A framework for animating open-domain still images by leveraging strong video diffusion priors, enabling realistic and dynamic motion synthesis for diverse visual content.
image-animationimage-to-videovideo-generation -
I2V-Adapter Discovered by AI
2023
A lightweight, training-free image-to-video adapter designed for diffusion models, enabling high-quality image-to-video generation without altering the original text-to-video model weights. It preserves the spatial structure and identity of the input image while generating temporally coherent motion.
adapterdiffusionimage-to-video -
I2VGen-XL Discovered by AI
2023
A cascaded image-to-video synthesis approach developed by Alibaba, designed to enhance semantic accuracy, detail continuity, and clarity in generated videos.
cascadeddiffusionimage-to-video -
LivePhoto Discovered by AI
2023
A practical system for real image animation that allows users to animate static images using text descriptions with precise motion intensity control.
diffusionimage-to-videomotion-control -
PIA Discovered by AI
2023
A personalized image animation framework that acts as a plug-and-play module for text-to-image models, allowing users to animate images while preserving high-fidelity details and specific styles.
image-to-videopersonalized-generationstyle-preservation -
SEINE Discovered by AI
2023
A short-to-long video diffusion model designed for generative transition and prediction, allowing high-quality image-to-video animation and seamless scene transitions.
image-to-videotransitionvideo-generation -
2023
An open-weights latent video diffusion model for image-to-video and short clip generation, a base for anime/game motion experiments and fine-tunes.
diffusionopen-weights
Proprietary / Industry Tools¶
-
Haiper Discovered by AI
2024
A commercial AI video generation platform that offers advanced text-to-video and image-to-video tools, designed to produce high-quality, stable animations and short-form video content.
commercialimage-to-videotext-to-video -
Haiper AI Discovered by AI
2024
A commercial AI video generation platform that allows users to create high-quality, realistic, or stylized videos from text prompts and static images.
commercialimage-to-videosaas -
Runway Gen-3 Alpha Discovered by AI
2024
A frontier proprietary video generation model by Runway that offers high-fidelity, consistent, and highly controllable text-to-video and image-to-video synthesis.
controlimage-to-videoproprietary -
Viggle Discovered by AI
2024
A popular proprietary AI video generation platform specializing in character animation and motion transfer. It allows users to animate static character images using text prompts or reference motion videos, widely used in game and anime production for rapid prototyping and meme creation.
character-animationmotion-transferproprietary -
Viggle AI Discovered by AI
2024
A popular proprietary AI animation platform that specializes in character motion transfer, allowing users to animate static character images using motion templates or custom video clips.
character-animationimage-to-videomotion-transfer
