Microsoft Patents a Faster Way to Draw Lines on 3D Surfaces in VR
Drawing a line on a bumpy 3D surface sounds simple, but every curve, ridge, or crater can make it a slow, expensive calculation. Microsoft's new patent describes a method that does the math for every part of the line at the same time, making real-time drawing on complex 3D objects practical on a headset.
What Microsoft's 3D mesh drawing tool actually does
You're wearing a mixed-reality headset, and you want to mark a route across a 3D map of rugged terrain. Every hill and valley means the line has to hug the surface, bending and dipping as you draw. On today's hardware, computing all those bends one step at a time can slow the whole experience to a crawl.
Microsoft's patent describes a drawing tool that handles all those bends at the same time, rather than one after another. The system looks at every triangle that makes up the 3D surface, figures out where your line crosses each one, and stitches those crossing points together into one smooth stroke. The result is a line that conforms to the surface shape, rendered quickly enough to feel responsive.
The company says this approach is much more efficient than the conventional method, called ray tracing, which fires a virtual beam from the user's pen and checks for collisions point by point. This matters most for complex objects like terrain maps or oddly shaped real-world things scanned into VR.
… excluding from rendering any triangle intersection segment having the two intersecting points outside the range; replacing any points of the triangle intersection segments falling outside the range with a nearer point selected from the starting point or the ending point …
Translation: The system trims the line so it stops precisely at the user defined start and end points.
How the parallel triangle-intersection method works
The patent describes a method for rendering a drawn line on a 3D mesh, which is what engineers call the web of tiny triangles that computers use to model any 3D surface, from a human face to a mountain range.
When a user draws a line in VR or mixed reality, the system:
- Identifies every triangle in the mesh that the drawn line's plane (an imaginary flat sheet cutting through the 3D object) intersects.
- For each of those triangles, connects the two points where the plane crosses the triangle's edges, creating a short triangle intersection segment.
- Trims any segments that fall outside the line's defined start-to-end range, replacing stray endpoints with the nearer endpoint of the actual stroke.
- Combines all the trimmed segments into a final set of rendering line segments made of basic geometric shapes, then draws the whole aggregated line at once.
The key engineering claim is that all triangles are processed in parallel, meaning the device's processor handles them simultaneously rather than sequentially. That's what makes it computationally efficient compared to ray tracing (a technique that checks collisions one ray at a time, which becomes slow when the mesh has thousands of triangles).
The patent targets devices like mixed-reality and virtual-reality headsets, where processing headroom is limited and users expect drawing to feel instant.
The parallel processing drawing method simultaneously computes each of the individual line segments to conform the virtually drawn line to the mesh. The method is computationally efficient compared to conventional ray tracing …
Translation: Processing the whole line at once instead of using ray tracing makes the drawing tool much faster.
What this means for VR annotation and spatial design
For anyone using a VR or mixed-reality headset to annotate, plan, or design in 3D space, lag in a drawing tool breaks the sense of presence. Microsoft's run of mixed-reality filings suggests the company sees spatial annotation as a core workflow, not a novelty feature. A method that makes freehand line-drawing on complex surfaces fast enough to feel natural could matter for fields like construction planning, military mapping, and surgical training, where marking up 3D models directly is genuinely useful.
The efficiency gain also has a hardware angle. Headsets have limited battery and processing power compared to desktop workstations. A drawing method that offloads less work and runs faster is the kind of optimization that keeps a tool usable on a standalone device rather than requiring a tethered PC.
Microsoft's seventh filing we've tracked in the AR glasses race since May builds on earlier applications covering a wider projected field and eye-tracked screen dimming.
Claim 1 is written broadly enough to cover almost any scenario where a user draws a line on a triangle-based 3D mesh on a computing device. The claim does not restrict itself to VR headsets, to any particular type of mesh, or to any specific parallel processing chip. That breadth means, if granted, the claim could apply to tablet apps, desktop 3D software, and game engines, not just Microsoft's own HoloLens or any future headset.
That breadth is also the claim's vulnerability. Triangle-mesh intersection math is well-established in computer graphics; the novelty Microsoft is staking out is the parallel, all-at-once approach combined with the specific trimming and aggregation steps. A patent examiner will likely look hard at whether those steps, in combination, are sufficiently distinct from existing parallel GPU rendering techniques that have been standard practice for years.
If the claim survives examination at its current width, it covers a lot of territory in 3D annotation software. If it gets narrowed, Microsoft probably retains protection for the specific trimming-and-aggregation pipeline, which is narrower but still practically useful for its own mixed-reality products.
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
41 drawing sheets from US 2026/0260426 A1 · click any drawing to enlarge
Want this weekly breakdown for a company we don't cover? Patentlyze Pro →