Mirror Polygons - compgeom - Coauthor
See Zach’s blog about Tokarski’s polygons with mirrored edges where illuminating one point keeps another point dark. But these examples crucially rely on light rays dying whenever they hit a vertex. Can we define a model of reflection (specifically at vertices) where we get a positive result? I raised the possibility that, if we allow lightrays to reflect from vertices according to some model, then perhaps we could obtain a positive result, e.g., a light source at any interior point illuminates the entire polygon—no dark points. Zach showed that the model I suggested allowed a lightray to bend around a reflex vertex, effectively ruling out that model. Here is a variation on the same idea, which at least avoids bending around. (1) Imagine a vertex 𝑣 v truncated orthogonal to the angle bisector. For 𝑣 v convex, lightrays reflect across the angle bisector. For 𝑣 v reflex, we again truncate orthogonal to the angle bisector. But now, if a ray heads toward 𝑣 v from below the line 𝑎
Coauthor
related reading
- Lights and Shadows – Bartosz Ciechanowskiciechanow.ski
- Visualizing Delaunay Triangulationianthehenry.com
- reflection - At what direction does a ray reflect at a 90 degrees corner? - Physics Stack Exchangephysics.stackexchange.com
- Illumination problem - Wikipediaen.wikipedia.org
- A point in many trianglesborisbukh.org
- [2309.09973] Monochromatic boxes of unit volumearxiv.org
- Convex and Concave Dispositionsvitalik.eth.limo
- Definability of Truth in Probabilistic Logic - Christiano 2013intelligence.org
- optics - Why do metals have high optical reflectivity? - Physics Stack Exchangephysics.stackexchange.com
- Polygon rectangulation, part 1: Minimum number of rectangles | Nanoexplanationsnanoexplanations.wordpress.com
- Inscribed square problem - Wikipediaen.wikipedia.org
- Differentiable Path Tracing on the GPU/TPUblog.evjang.com