Samsung Patents a Camera That Remembers Who You Were Filming
Samsung is working on a camera system that tags the subjects in your video while you record, then uses those tags later to pull up every photo or clip containing that same person or object.
How Samsung's subject-tagging video system works
Ever tried to find that one clip of your kid at the school play, buried somewhere in two years of phone videos? That search usually means scrolling forever or hoping your memory is better than your filing system.
Samsung's patent describes a camera that watches what you're filming and silently attaches notes to the recording. Those notes describe the subjects in the frame, so later, when you tap on a person or object in any clip, the phone can scan your entire library and surface every other video or photo that includes that same subject.
The system also lets you reframe a wide-angle video after the fact. While you're recording with the full camera view, you can tap a person or object on screen, and the phone adjusts the displayed shot to focus on them, while still saving the full original footage in the background.
based on a user input, identify a visual object corresponding to a subject of a media content; based on metadata associated with a plurality of media contents accessible by the electronic device, identify, from among the plurality of media contents, one or more images including a visual object corresponding to the subject …
Translation: The camera recognizes who is in the frame and searches past photos for the same person.
How the metadata links subjects across your entire library
The patent covers two connected ideas that work together around a concept called field-of-view (FoV), which just means how wide or narrow the camera's view is at any moment.
First, while recording, the camera captures a wide first FoV. If you tap a subject on the preview screen, the phone shifts the display to a tighter second FoV centered on that subject. Crucially, it keeps recording the full wide shot and writes metadata (small data tags attached to the file, invisible to you) that remembers which FoV was active and which subject you selected at each moment.
Second, and this is the broader claim in the patent, the device can use that metadata to do a library-wide search. The system:
- Identifies a visual object or subject in a piece of media you select
- Scans metadata attached to every accessible photo and video
- Pulls out every file that contains a matching subject
- Plays those results back as a continuous video
The metadata does the heavy lifting here. Because the tags are written at recording time rather than analyzed retroactively, the search can theoretically run faster and work even on files that were recorded in different sessions or on different days.
An electronic device receives a shooting input while displaying a preview image based on at least a portion of images obtained through a camera having a first field-of-view (FoV).
Translation: The device takes a video while framing a wider scene through the camera lens.
What this means for Galaxy video search
For most Galaxy users, the practical payoff would be a photo-and-video search that actually works on your own terms rather than requiring you to remember a date or album name. Tap on your dog in one clip, and the phone finds every clip of your dog. That kind of subject-driven browsing is something Google Photos already attempts using on-device AI face grouping, so Samsung is clearly trying to build a comparable system at the camera level rather than the gallery level.
The reframing angle is also interesting for anyone who shoots events. Recording wide and selecting a tighter subject view after the fact means you never have to choose between capturing the whole scene and getting a close-up shot. Samsung's steady filing work in camera and display intelligence suggests this is part of a longer effort to bring those post-production decisions into the recording moment itself.
That makes this Samsung's 110th filing we've tracked since May in our camera sensor push, adding to work like the flat AI lens and the two-stage photo sharpener.
The cameras already exist. This patent is entirely about software: while you film, the phone silently notes which people or objects appeared on screen, so you can jump straight to those moments later instead of scrubbing through an hour of footage.
The trickiest piece is doing that recognition live, on a phone, without killing the battery. A practical first version would likely skip recognizing strangers entirely and only track faces the phone already knows from your photos, triggered only when you tap the screen on someone.
That narrower version could ship on phones available today, which makes this closer to a real feature than most patents suggest.
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
11 drawing sheets from US 2026/0270550 A1 · click any drawing to enlarge
Want this weekly breakdown for a company we don't cover? Patentlyze Pro →