Browsing by Author "Manzi, Marco"

Now showing 1 - 4 of 4
  • Thumbnail Image
    Item
    Automatic Feature Selection for Denoising Volumetric Renderings
    (The Eurographics Association and John Wiley & Sons Ltd., 2022) Zhang, Xianyao; Ott, Melvin; Manzi, Marco; Gross, Markus; Papas, Marios; Ghosh, Abhijeet; Wei, Li-Yi
    We propose a method for constructing feature sets that significantly improve the quality of neural denoisers for Monte Carlo renderings with volumetric content. Starting from a large set of hand-crafted features, we propose a feature selection process to identify significantly pruned near-optimal subsets. While a naive approach would require training and testing a separate denoiser for every possible feature combination, our selection process requires training of only a single probe denoiser for the selection task. Moreover, our approximate solution has an asymptotic complexity that is quadratic to the number of features compared to the exponential complexity of the naive approach, while also producing near-optimal solutions. We demonstrate the usefulness of our approach on various state-of-the-art denoising methods for volumetric content. We observe improvements in denoising quality when using our automatically selected feature sets over the hand-crafted sets proposed by the original methods.
  • Thumbnail Image
    Item
    Deep Compositional Denoising for High-quality Monte Carlo Rendering
    (The Eurographics Association and John Wiley & Sons Ltd., 2021) Zhang, Xianyao; Manzi, Marco; Vogels, Thijs; Dahlberg, Henrik; Gross, Markus; Papas, Marios; Bousseau, Adrien and McGuire, Morgan
    We propose a deep-learning method for automatically decomposing noisy Monte Carlo renderings into components that kernelpredicting denoisers can denoise more effectively. In our model, a neural decomposition module learns to predict noisy components and corresponding feature maps, which are consecutively reconstructed by a denoising module. The components are predicted based on statistics aggregated at the pixel level by the renderer. Denoising these components individually allows the use of per-component kernels that adapt to each component's noisy signal characteristics. Experimentally, we show that the proposed decomposition module consistently improves the denoising quality of current state-of-the-art kernel-predicting denoisers on large-scale academic and production datasets.
  • Thumbnail Image
    Item
    Deep Compositional Denoising on Frame Sequences
    (The Eurographics Association, 2023) Zhang, Xianyao; Röthlin, Gerhard; Manzi, Marco; Gross, Markus; Papas, Marios; Ritschel, Tobias; Weidlich, Andrea
    Path tracing is the prevalent rendering algorithm in the animated movies and visual effects industry, thanks to its simplicity and ability to render physically plausible lighting effects. However, we must simulate millions of light paths before producing one final image, and error manifests as noise during rendering. In fact, it can take tens or even hundreds of CPU hours on a modern computer to render a plausible frame in a recent animated movie. Movie production and the VFX industry rely on image-based denoising algorithms to ameliorate the rendering cost, which suppresses the noise due to rendering by reusing information in the neighborhood of the pixels both spatially and temporally.
  • Thumbnail Image
    Item
    Robust Image Denoising using Kernel Predicting Networks
    (The Eurographics Association, 2021) Cai, Zhilin; Zhang, Yang; Manzi, Marco; Oztireli, Cengiz; Gross, Markus; Aydin, Tunç Ozan; Theisel, Holger and Wimmer, Michael
    We present a new method for designing high quality denoisers that are robust to varying noise characteristics of input images. Instead of taking a conventional blind denoising approach or relying on explicit noise parameter estimation networks as well as invertible camera imaging pipeline models, we propose a two-stage model that first processes an input image with a small set of specialized denoisers, and then passes the resulting intermediate denoised images to a kernel predicting network that estimates per-pixel denoising kernels. We demonstrate that our approach achieves robustness to noise parameters at a level that exceeds comparable blind denoisers, while also coming close to state-of-the-art denoising quality for camera sensor noise.