skeleton-tracing

Why we included this project

Most skeletonization tools stop at a raster result, which means you still have to trace the strokes yourself if you want vector geometry. This library skips that step: it runs a classic Zhang-Suen thinning pass, then a divide-and-conquer tracing stage that splits the image, processes each half, and merges the segments back into ordered polylines. Because the output is a set of single strokes rather than enclosing outlines, it suits handwriting recognition, drawing tools, and other shape-analysis work where you need clean centerline geometry. The implementation is available in C, C++, Python, JavaScript, Rust, Go, and a dozen more languages, so it slots into most existing stacks without much friction, and the online demo plus benchmarks give you a quick way to judge whether the real-time performance holds up for your workload.