University of California

Nima Khademi Kalantari

Pradeep Sen

Steve Bako

Ravi Ramamoorthi

Devices, systems and methods for generating high dynamic range images and video from a set of low dynamic range images and video using convolution neural networks (CNNs) are described. One exemplary method for generating high dynamic range visual media includes generating, using a first CNN to merge a first set of images having a first dynamic range, a final image having a second dynamic range that is greater than the first dynamic range. Another exemplary method for generating training data includes generating sets of static and dynamic images having a first dynamic range, generating, based on a weighted sum of the set of static images, a set of ground truth images having a second dynamic range greater than the first dynamic range, and replacing at least one of the set of dynamic images with an image from the set of static images to generate a set of training images.

Generation of high dynamic range visual media

A method of rendering an image includes Monte Carlo rendering a scene to produce a noisy image. The noisy image is processed to render an output image. The processing applies a machine learning model that utilizes colors and/or features from the rendering system for denoising the noisy image and/or to for adaptively placing samples during rendering.

Growing community of inventors

La Jolla, CA, United States of America

Nima Khademi Kalantari