Intel introduced the first experimental release of the oidn library (Open Image Denoise), which offers a collection of filters for eliminating noise in images prepared using ray tracing based rendering systems. The library is being developed as part of a larger Intel Rendering Framework project aimed at developing software for visualizing scientific calculations (SDVis (Software Defined Visualization)), which includes Embree ray tracing library, GLuRay photorealistic rendering system, OSPRay distributed ray tracing system and OpenSWR software rasterization system. Code written in C ++ and published under the Apache 2.0 license.
The goal of the Open Image Denoise library is to provide high-quality, efficient and easy-to-use noise reduction functions that can be used to improve the quality of the ray tracing results. The proposed filters allow, based on the result of the reduced ray tracing cycle, to obtain a final quality level comparable to the result of a more costly and lengthy process of detailed rendering.
The proposed algorithms provide screening of random noise, simulated by the Monte Carlo method, characteristic of stochastic ray tracing algorithms, such as path-based rendering. To achieve high quality rendering in such algorithms, tracing of a very large number of rays is required, otherwise noticeable artifacts in the form of random noise appear on the resulting image.
The use of Open Image Denoise allows you to reduce the number of necessary calculations by several orders of magnitude when calculating each pixel. As a result, it is possible to generate much initially noisy images and bring it to an acceptable quality using fast noise reduction algorithms. Depending on the equipment used, this approach can even be used for interactive ray tracing and imaging on the run.