Kernel-predicting convolutional networks for denoising Monte Carlo renderings

351 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Omni-Dimensional Dynamic Convolution (2022, neutral): “Kernel Prediction Networks (Bako et al., 2017; Mildenhall et al., 2018) leverage a CNN architecture to predict spatially varying kernels used for video denoising.”
Pixel-Adaptive Convolutional Neural Networks (2019, background): “These are called “Dynamic Filter Networks” (DFN) [47, 22, 12, 46] (also referred to as cross-convolution [47] or kernel prediction networks [4]) and have been shown to be useful in several computer vision tasks.”
Deep Learning on Image Denoising: An overview (2019, neutral): “For non-linear noisy images, CNNs with kernel methods proved useful [14, 235].”
Burst Denoising with Kernel Prediction Networks (2017, methodology): “[2] we do not normalize the predicted filters with a softmax, thereby allowing predicted kernels to have negative values.”
Deep blending for free-viewpoint image-based rendering (2018, methodology): “Similarly to [Bako et al. 2017], we observe that directly regressing the output image works well, but network convergence is much slower.”