Samsung · Filed Mar 25, 2026 · Published Aug 6, 2026 · verified — real USPTO data

Samsung Patents a Multi-Camera System That Eliminates Video Shake

Most phone stabilization fights shake after the fact. Samsung's new patent tries something different: pulling the steadiest slice from every camera at once and stitching those slices together into one rock-solid frame.

Samsung Patent: Hyper-Stabilized Video From Multiple Cameras — figure from US 2026/0230687 A1
Figure from the official USPTO publication.
See all 5 drawings from this filing ↓
Publication number US 2026/0230687 A1
Applicant SAMSUNG ELECTRONICS CO., LTD.
Filing date Mar 25, 2026
Publication date Aug 6, 2026
Inventors Ritesh SINGH
CPC classification 725/116
Grant likelihood Medium
Examiner CENTRAL, DOCKET (Art Unit OPAP)
Status Docketed New Case - Ready for Examination (May 22, 2026)
Parent application is a Continuation of PCTKR2024016445 (filed 2024-10-25)
Document 19 claims

What Samsung's multi-camera stabilization actually does

Imagine you're filming a friend's birthday toast while holding your phone with one hand. Even with optical stabilization on, the footage comes out a little wobbly. Traditional stabilization tries to smooth that out after the fact, but it's still working from one shaky source.

Samsung's patent describes a different approach. Your phone already has multiple cameras pointing at roughly the same scene. This system looks at the footage from all of them at once, figures out the steadiest central region each camera can reliably capture given its angle and zoom, and then cuts out that steady slice from each camera's feed.

Those slices are then assembled into a single composite video frame. The result is a final video that draws only from the parts of each camera's picture that barely moved at all, giving you footage that's far steadier than any one camera could deliver on its own.

How the system picks and stitches the steadiest frame regions

The patent describes a process that runs across the multi-camera array of a device like a smartphone. Here's the sequence:

  • Field width and tilt detection: The system reads each camera's field of view (how wide an angle it captures) and its tilt angle (how much the camera is physically angled relative to the others).
  • Largest Central Square Field of View (LCSFoV): For each camera, the system calculates the biggest square crop that sits at the center of the frame and remains usable no matter how the device tilts. Think of it as the guaranteed-steady bullseye of each lens.
  • Retention area: Across all cameras, the system finds a common retention area, the region that all cameras can reliably cover within their individual LCSFoVs.
  • Fragment extraction and assembly: It pulls the matching fragment from each camera's video frames and stitches those fragments together into one composite output frame.

The approach is different from software stabilization (which crops and shifts a single camera's frames after recording) and from optical image stabilization (which physically moves a lens element). Instead, it treats the multi-camera array as a single distributed sensor and picks only the steadiest data from each.

We find one patent like this every day. Get the best of each week in your inbox, free →

What this means for Galaxy video recording

Samsung's Galaxy phones already carry three or four cameras, and right now each lens mostly competes for the user's attention rather than cooperating with the others. This patent points toward a future where those cameras act as a team during recording, not just at the moment you switch between lenses.

For you as a user, the promise is video that stays steady even when you're moving, without the telltale zoomed-in, cropped look that aggressive software stabilization produces today. It could also matter for action or sports recording, where shake is unavoidable and every fraction of a degree of tilt costs you image quality.

Editorial take

This is a genuinely interesting idea because it reframes what a multi-camera phone array is actually for. Most manufacturers use multiple cameras to give you zoom options or depth data; using them cooperatively to build one steadier frame is a different bet entirely. Whether the stitching artifacts are manageable in real footage is the engineering question that will decide whether this goes anywhere.

There are more where this came from

We read every patent application Big Tech publishes and send you the ones worth knowing. Plain English, free, every week.

The drawings

5 drawing sheets from US 2026/0230687 A1 · click any drawing to enlarge

Patent filing page

Source. Full patent text and figures from the official USPTO publication PDF.

Editorial commentary on a publicly published patent application. Not legal advice.