This product is a novel denoising diffusion probabilistic model that learns to sample from an unobserved signal distribution instead of directly observing it. It measures samples from the known differentiable forward model. It can directly sample from a partially observed unknown signal distribution and is suitable for computer vision tasks. In inverse graphics, it can generate a 3D scene distribution consistent with a single 2D input image. The product offers flexible pricing and targets the image processing and computer vision domains.