StippleDiffusion: Capacity-Constrained Stippling using Controlled Diffusion

Name: StippleDiffusion: Capacity-Constrained Stippling using Controlled Diffusion
Uploaded: 2026-10-22T06:23:33.913109+00:00
Duration: 589 s
Description: StippleDiffusion is a diffusion-based sampler that creates stipple patterns by simultaneously learning a local point distribution and enforcing a continuous, image-defined capacity constraint at inference. It achieves high-quality, spatially controlled stippling using a single trained model that works for any target density without needing per-density optimization.

StippleDiffusion is a diffusion-based sampler that creates stipple patterns by simultaneously learning a local point distribution and enforcing a continuous, image-defined capacity constraint at inference. It achieves high-quality, spatially controlled stippling using a single trained model that works for any target density without needing per-density optimization.

Open the full episode