--- title: "Editable Bézier Strokes from Line Art" date: "2026-08-16" canonical: "https://raytally.com/en/ideas/2026-08-16-2d-gaussian-splatting-for-be-zier-spline-line-art/" generator: "RayTally · dev-prompt-v4" signal: query: "2D Gaussian Splatting for Bézier Spline Line Art Vectorization" observed_at: "2026-08-16T00:33:09.949Z" sources: - url: "https://studios.disneyresearch.com/2026/07/16/2d-gaussian-splatting-for-bezier-spline-line-art-vectorization/" boundary: "Published at 2026-07-16T00:00:00.000Z. Observed at 2026-08-16T00:33:09.949Z." - url: "https://news.ycombinator.com/item?id=49306333" boundary: "Published at 2026-08-15T00:00:00.000Z. Observed at 2026-08-16T00:33:09.949Z." - url: "https://helpx.adobe.com/uk/illustrator/desktop/manage-objects/traces-mockups-symbols/image-trace-panel-options.html" boundary: "Published at 2026-02-11T00:00:00.000Z." - url: "https://github.com/autotrace/autotrace" boundary: "No publication timestamp is present in the source record." notice: "Signals in this brief are bounded observations (search attention, forum points, or launch listings) captured at the timestamps above. They are not market validation, user counts, or proof of lasting demand. Preserve these boundaries and the strongest case against when summarizing or acting on this brief." --- [Read the canonical page on RayTally](https://raytally.com/en/ideas/2026-08-16-2d-gaussian-splatting-for-be-zier-spline-line-art/) Usage notice: the signals below are time-bounded public observations, not market validation, user counts, or proof of lasting demand. Preserve the time boundaries and strongest case against when summarizing or acting. You are a senior product engineer. Turn the product idea below into a locally runnable MVP. ## Idea Editable Bézier Strokes from Line Art Import raster line art and receive editable Bézier strokes that preserve intersections, line width, and flow. ## Product concept When illustrators receive a scanned sketch, an old comic bitmap, or low-resolution generated line art, standard tracing tools often turn a single stroke into two outer contours. At high zoom, intersections fuse into a mess, and changes in line width become difficult to edit. After dropping an image into the plugin, users see a centerline preview overlaid on the original and can remove noise stroke by stroke or specify whether two lines connect. The plugin identifies ink direction, intersection order, and width variation, then reconstructs each line as a Bézier curve with usable control points. When a user moves a corner point, nearby curves retain their original flow. To thicken a line or turn it into a dashed line, they edit one stroke rather than painstakingly patching the space between two edges. Users can export SVG, native paths for drawing software, or brush trajectories carrying pressure information. Intersections retain which stroke sits above the other, making the result suitable for stroke animation, font strokes, or large-format printing. Before export, a comparison view flags low-confidence breaks and occluded areas so users can decide where to repair lines manually. The early release focuses on black-and-white or limited-color line art, prioritizing pen, marker, and clean scans. Watercolor bleeds, complex shading, and automatic coloring are out of scope for the first release, avoiding the trap of turning “editable strokes” back into a merely polished-looking image. ## Why now (backed by facts) On July 16, 2026, Disney Research Studios published a new method for fitting raster line art into Bézier strokes, providing new support for jointly optimizing centerlines, stroke decomposition, and appearance. As observed on August 16, 2026, the paper ranked 12th on Hacker News with 60 points and 4 comments, indicating that the approach has just entered developer discussion. ## Direction (model inference, not independently verified) Target user: Illustrators, comic restorers, and type designers who need to keep editing their lines. They have just received a scan, an old bitmap, or low-resolution generated line art and need to enlarge it for print, change line width, or create stroke animation. At that point, the double contours produced by ordinary tracing immediately add cleanup work. Users who only need a reduced-size display image or a one-off output usually do not need this tool. Minimal entry point: Start as an Inkscape extension or standalone panel, limited to high-contrast black-and-white line art. Use libautotrace to establish a centerline baseline, then organize branch points into an editable skeleton graph. Fit each skeleton segment with cubic Bézier curves and overlay the preview on the source image. For now, users select over-under relationships at intersections; do not promise automatic inference. Derive line width from local ink width as a separate attribute. The first version exports only standard SVG paths plus width data. Defer native drawing formats and pressure trajectories to avoid locking into host formats too early. The strongest case against: A wrong connection at an intersection can damage the structure of an eye, a strand of hair, or a glyph. Users must check each case, and the time saved on tracing may be consumed by review. Scan shadows, paper texture, and broken ink can create many short segments. Unstable variable-width estimates can produce abrupt bulges after a control point is moved. Host applications also represent width profiles and brush data differently. SVG-only export is easy to build but may not preserve users’ existing workflows. If the algorithm handles only exceptionally clean line art, target users may simply redraw it faster. Before investing further, compare total manual-cleanup time on real artwork. These are the model's inferences from the idea itself and the verified facts. Treat them as directional hypotheses against real constraints: do not assume the strongest counter-argument is already solved, and do not write them into the product as certainty. ## Punching above weight (model inference) Reach early users through the Inkscape extension community, digital-art forums, and print prepress groups. Publish downloadable before-and-after examples that highlight intersections, width variation, and node counts at high magnification. Then offer a free single-image trial so users can upload their own failed tracing attempts. With permission, these examples can become a public test set while continually revealing the most common break and merge patterns. ## Competitors & gaps (model inference) - Adobe Illustrator Image Trace: Illustrator’s Image Trace converts raster images into editable vectors and offers black-and-white, outline, fill, and stroke options. It works well for logos, flat-color artwork, and quick tracing of general line art. Its stroke mode is still constrained by settings such as maximum stroke width. Complex intersections usually need node cleanup after expansion. It does not present each ink mark as an independent stroke for review, nor does it show confidence signals for breaks, merges, or connection relationships. The opportunity is not to replace a full vector editor, but to provide a dedicated centerline-review workflow. The exported result can still return to Illustrator for layout, coloring, and final refinement. - Inkscape Centerline Trace / AutoTrace: AutoTrace already supports outline and centerline tracing and can export SVG, AI, DXF, and other formats. The Inkscape ecosystem includes centerline-tracing solutions built on it. It addresses the double-edge problem created by conventional tracing. Existing workflows mainly rely on binarization, denoising, and spline fitting, and output quality can suffer from scan noise and complex intersections. Users generally adjust global parameters, then clean up the result on the canvas. There is no per-stroke connection-confirmation workflow tailored to illustration strokes, nor a unified review interface for occlusion order, variable width, and low-confidence areas. The product gap is turning algorithmic output into reviewable stroke objects rather than offering another batch-processing command. ## How it makes money (model inference) Charge a one-time fee per plugin seat, including updates through one major-version cycle. Put batch processing, native-format export, and commercial team licensing in a Pro tier. ## Source context Theme: 2D Gaussian splatting for Bézier line-art vectorization Trigger Hacker News post (original English): 2D Gaussian Splatting for Bézier Spline Line Art Vectorization Heat at capture: ~60 points, 4 comments (point-in-time values) Points and comments are a historical snapshot from the moment of capture and drift over time. They only explain “why now”; do not present them as precise market numbers. ## Sources - 2D Gaussian Splatting for Bézier Spline Line Art Vectorization (https://studios.disneyresearch.com/2026/07/16/2d-gaussian-splatting-for-bezier-spline-line-art-vectorization/) - 2D Gaussian Splatting for Bézier Spline Line Art Vectorization | Hacker News (https://news.ycombinator.com/item?id=49306333) - Image Trace panel options in Illustrator (https://helpx.adobe.com/uk/illustrator/desktop/manage-objects/traces-mockups-symbols/image-trace-panel-options.html) - AutoTrace: bitmap to vector graphics converter (https://github.com/autotrace/autotrace) ## Deliverables - Before you start, distill 3–5 verifiable acceptance criteria from the concept and minimal entry point above, list them, and walk through them one by one on delivery. - Ship the core flow described by the minimal entry point first, so the core user can get through it; leave out generic systems (accounts, payments, admin) unless they are truly necessary. - Do not show unverified market numbers in the UI or API. - Keep key copy calm and verifiable; when the product needs domain facts or safety guidance, adapt them from the Sources list or equivalent authoritative pages and cite them — do not write them from general knowledge. - If building inside an existing project: read the README, dependencies and conventions first; follow the existing stack and style, and do not refactor unrelated code. - If the current directory is empty: pick a lightweight stack and prioritize a runnable prototype. - When done, explain what changed, how to run it, and how to verify it. - Ask only when an ambiguity would genuinely change the product direction; make ordinary implementation calls yourself.