Have you ever tried to draw a perfect circle or ellipse by hand? It’s not easy, right? That’s why I’ve been working on a library that can approximate geometric shapes from a sequence of 2D points. The goal is to take a list of (x, y) points and fit the best-matching shape, whether it’s a circle, ellipse, triangle, or something more complex.
The possibilities are endless. Imagine using this tech for hand-drawn sketches, geometric recognition, or shape fitting in graphics and vision tasks. But to make it work, I need to test and improve the library using real-world or benchmark datasets.
That’s where you come in. I’m on the hunt for datasets that include:
* Point sequences or stroke data (like hand-drawn shapes)
* Annotated datasets where the intended shape is known
* Noisy samples that simulate real drawing or sensor data
If you know of any existing datasets that fit the bill, I’d love to hear about them. You can check out my library on GitHub to see what I’m working on:
Together, we can create a tool that brings precision and accuracy to shape approximation.