CameraAnything: Refilming Videos with Arbitrary Camera Control

2026-07-27

AI Summary

Existing video editing methods struggle to disentangle and jointly control intrinsic camera parameters (e.g., focal length, resolution) and extrinsic parameters (e.g., viewpoint), often relying on costly 3D reconstruction or supporting only limited adjustments. This work proposes CameraAnything, the first framework enabling unified, controllable video re-rendering of both intrinsic and extrinsic camera properties. By injecting pixel-wise Plücker rays and employing resolution-aware 3D RoPE positional encoding in latent space, our method jointly models camera pose, focal length, and native resolution without cropping or outpainting. We introduce a scalable multi-camera synthetic data pipeline and an orthogonal training strategy to mitigate the scarcity of paired real-world data. CameraAnything supports arbitrary viewpoint changes, focal length adjustment, resolution adaptation, and smooth multi-shot transitions within a single generation pass, demonstrating high flexibility and practicality for film production and cross-platform content adaptation.

Abstract

We introduce CameraAnything, the first unified framework for camera controlled video editing that enables joint control of both intrinsic and extrinsic camera parameters. Existing approaches either rely on expensive 3D reconstruction to achieve full camera functionality or restrict editing to extrinsic parameter manipulation. Moreover, the coupled influence of intrinsic and extrinsic parameters on video appearance makes disentangled modeling particularly challenging. To address this, we adopt per-pixel Plücker ray injection alongside resolution-aware 3D RoPE in self-attention, building both camera conditioning and spatial positional encoding on the target latent to jointly control camera position, focal length, and native resolution editing without cropping or outpainting. To overcome the scarcity of paired training data, we further develop a scalable synthetic pipeline that constructs diverse dynamic scenes through structured multi-camera recording and generates synchronized videos with varied camera configurations. With a tailored orthogonal training strategy, CameraAnything enables expressive video reshooting with arbitrary viewpoint control, focal length adjustment, resolution adaptation, and multi-shot transitions within a single generation process, offering strong practical value for cinematic video editing and cross-platform content adaptation in video production.

Problem

• camera control
• video editing
• intrinsic parameters
• extrinsic parameters
• 3D reconstruction

Innovation

• camera control
• Plücker ray injection
• 3D RoPE
• video editing
• synthetic data pipeline