Image-to-Video Diffusion: From Foundations to Open Frontiers

2026-05-17

AI Summary

Image-to-video (I2V) generation faces core challenges in content consistency, identity preservation, and motion coherence, yet lacks a systematic survey and unified taxonomy. This work establishes diffusion-based I2V generation as a distinct research direction and introduces a dedicated classification framework, identifying four key design components: condition encoding, temporal modeling, noise prior design, and spatiotemporal upsampling. By systematically organizing existing approaches according to model architectures and training paradigms, the paper clarifies task formulations, datasets, evaluation metrics, and methodological advances. It further delineates critical technical pathways, representative application scenarios, and open challenges, thereby promoting standardization and systematic progress in the field.

Abstract

Diffusion-based \textit{image-to-video} (I2V) generation has become a central direction in generative models by turning a reference image, with optional conditions, into a temporally coherent video. Compared with broader video generation settings, this task places stricter demands on content consistency, identity preservation, and motion coherence. Although the literature grows rapidly, existing works mostly discuss I2V generation within broader topics and still lack a dedicated taxonomy together with a systematic analysis centered on this field. This work addresses that gap by treating diffusion I2V generation as a standalone subject. It first reviews the task formulation, model architectures, datasets, and evaluation metrics, and then organizes existing methods through a taxonomy based on architecture and training paradigm. It further distills four core designs, namely condition encoding, temporal modeling, noise prior design, and spatial-temporal upsampling, and discusses representative application scenarios together with major open challenges.

Problem

• image-to-video
• diffusion models
• generative models
• temporal coherence
• content consistency

Innovation

• image-to-video
• diffusion models
• temporal coherence
• condition encoding
• taxonomy