Single-Camera Rolling Shutter Correction Using Strabismic Imaging

Researchers at Stanford have developed a novel method to correct motion-induced image distortion in rolling shutter cameras using a single camera sensor.

Rolling shutter cameras are widely used across consumer electronics, medical imaging, and autonomous systems. Rather than capturing an entire frame simultaneously, they read out image data sequentially, row by row, which introduces motion artifacts whenever the camera or subject moves. Existing solutions each carry meaningful limitations: motion-sensing hardware only compensates for camera movement, not a moving subject; software-based correction using machine learning can fail unpredictably on unfamiliar scenes; and upgrading to an alternative sensor architecture is often impractical at scale.

This technology applies strabismic imaging, a technique that uses two slightly offset, synchronized image captures, to enable motion correction from a single camera sensor. No machine learning is required. The correction is performed by deterministic mathematical algorithms that are broadly applicable across scene types and content, making the approach well-suited for integration into a wide range of imaging products.

Stage of Development: Proof of concept.