Publications

We present QW-Net, a novel 4-bit quantized network that performs temporally amortized supersampling for high-dynamic-range rendering.

We visualize – and use artificial neural networks to learn – the shapes of statistical sample distributions in Monte Carlo rendering, and apply that knowledge to produce improved (denoised) images from small sample sets.

We present Deep Illumination, a novel machine learning technique for approximating global illumination (GI) in real-time applications using a Conditional Generative Adversarial Network.