Hi! We have developed a tracing algorithm that blends Deep Learning, classical Machine Learning, and hand-crafted computer vision algorithms to turn regular bitmaps (JPEG, PNG, ...) into scalable vector graphics (SVG).
It does quite well in general, and especially on small-ish, blurry, noisy and otherwise degraded images that vectorizers normally choke on.
