AIAA is a techinique that uses a neural network to simulate the high-quality antialiasing of SSAA4x as a fast post process.
The model is currently trained on a set of renders of demo_scene at both stadard and supersampled resolutions downscaled for antialiasing.
Limited training data also confines this model's accuracy to demo_scene.
There is no official implementation yet. For now, clone the repository and run main.py which will walk you through generating data, setting model parameters, training, and running.
Run check_version.py to check Cuda Compatability.
Below is a comparison of a frame aliased, supersampled, and AIAA at 1080p (Click to view images in full resolution).
Alias
SSAA4x
AIAA
- Cuda-capable GPU with compute > 7.0 (RTX 20-Series or Newer)
- Python 3.9 - 3.12
- Pillow
- Torch 2.8.0 with Cuda 12.6
- Torchvision 0.23.0
- TQDM
- Triton-Windows 3.4.0.post20
- Added adjustable model parameters and automated execution