Asymmetric Loss Functions

Thijs Vogels, Fabrice Rousselle, Brian McWilliams, Gerhard Röthlin, Alex Harvill, David Adler, Mark Meyer, Jan Novák

Monte Carlo path tracing

Basic single-frame denoiser (Bako et al. 2017)

Kernel

Image

Color

© Disney

- Multi-renderer support source-aware encoders
- Temporal stability cross-frame denoising
- Denoising of low frequencies a multi-scale approach
- Adaptive sampling the error predictor
- User control asymmetric loss functions

- Multi-renderer support source-aware encoders
- Temporal stability cross-frame denoising
- Denoising of low frequencies a multi-scale approach
- Adaptive sampling the error predictor
- User control asymmetric loss functions

Baseline generalization ability

Renderer B Input

Trained on Renderer A data

Retrained on Renderer B data

Scene by nacimus (Blendswap)

Source-aware encoders

Data efficiency of encoder training

- Multi-renderer support source-aware encoders
- Temporal stability cross-frame denoising
- Denoising of low frequencies a multi-scale approach
- Adaptive sampling the error predictor
- User control asymmetric loss functions

Reusing a pre-trained single-frame denoiser

(Chaitanya et al. 2017)

- Multi-renderer support source-aware encoders
- Temporal stability cross-frame denoising
- Denoising of low frequencies a multi-scale approach
- Adaptive sampling the error predictor
- User control asymmetric loss functions

© Disney

Cheaper denoising of low frequencies

© Disney

- Multi-renderer support source-aware encoders
- Temporal stability cross-frame denoising
- Denoising of low frequencies a multi-scale approach
- Adaptive sampling the error predictor
- User control asymmetric loss functions

with an error predictor

- Multi-renderer support source-aware encoders
- Temporal stability cross-frame denoising
- Denoising of low frequencies a multi-scale approach
- Adaptive sampling the error predictor
- User control asymmetric loss functions

© Disney / Pixar

- Multi-renderer support source-aware encoders
- Temporal stability cross-frame denoising
- Denoising of low frequencies a multi-scale approach
- Adaptive sampling the error predictor
- User control asymmetric loss functions