Adobe, Inc.

Google Inc.

Michael Gharbi

Federico Perazzi

Zhihao Xia

Elya Shechtman

Kalyan Krishna Sunkavalli

Matthew David Fisher

Jonathan Tilton Barron

Connelly Barnes

Kevin Wampler

He Zhang

Jonathan Eisenmann

Samuel William Hasinoff

Jiawen Chen

Michal Lukác

Krishna Bhargava Mullia Lakshminarayana

Ke Wang

Camille Biscarrat

Mustafa Isik

Utkarsh Singhal

Chi Cheng Hsu

An image processing system employs a parametric model for image harmonization of composite images. The parametric model employs a two-stage approach to harmonize an input composite image. At a first stage, a color curve prediction model predicts color curve parameters for the composite image. At a second stage, the composite image with the color curve parameters is input to a shadow map prediction model, which predicts a shadow map. The predicted color curve parameters and shadow map are applied to the composite image to provide a harmonized composite image. In some aspects, the color curve parameters and shadow map are predicted using a lower-resolution composite image and up-sampled to apply to a higher-resolution version of the composite image. The harmonized composite image can be output with the predicted color curve parameters and/or shadow map, which can be modified by a user to further enhance the harmonized composite image.

Parametric composite image harmonization

The present disclosure relates to systems, methods, and non-transitory computer readable media that utilize a continuous kernel neural network that learns continuous reconstruction kernels to merge digital image samples in local neighborhoods and generate enhanced digital images from a plurality of burst digital images. For example, the disclosed systems can utilize an alignment model to align image samples from burst digital images to a common coordinate system (e.g., without resampling). In some embodiments, the disclosed systems generate localized latent vector representations of kernel neighborhoods and determines continuous displacement vectors between the image samples and output pixels of the enhanced digital image. The disclosed systems can utilize the continuous kernel network together with the latent vector representations and continuous displacement vectors to generated learned kernel weights for combining the image samples and generating an enhanced digital image.

Modeling continuous kernels to generate an enhanced digital image from a burst of digital images

A plurality of pixel-based sampling points are identified within an image, wherein sampling points of a pixel are distributed within the pixel. For individual sampling points of individual pixels, a corresponding radiance vector is estimated. A radiance vector includes one or more radiance values characterizing light received at a sampling point. A first machine learning module generates, for each pixel, a corresponding intermediate radiance feature vector, based on the radiance vectors associated with the sampling points within that pixel. A second machine learning module generates, for each pixel, a corresponding final radiance feature vector, based on an intermediate radiance feature vector for that pixel, and one or more other intermediate radiance feature vectors for one or more other pixels neighboring that pixel. One or more kernels are generated, based on the final radiance feature vectors, and applied to corresponding pixels of the image, to generate a lower noise image.

Denoising images rendered using Monte Carlo renderings

Certain embodiments involve techniques for efficiently estimating denoising kernels for generating denoised images. For instance, a neural network receives a noisy reference image to denoise. The neural network uses a kernel dictionary of base kernels and generates a coefficient vector for each pixel in the reference image such that the coefficient vector includes a coefficient value for each base kernel in the kernel dictionary, where the base kernels are combined to generate a denoising kernel and each coefficient value indicates a contribution of a given base kernel to a denoising kernel. The neural network calculates the denoising kernel for a given pixel by applying the coefficient vector for that pixel to the kernel dictionary. The neural network applies each denoising kernel to the respective pixel to generate a denoised output image.

Kernel prediction with kernel dictionary in image denoising

Embodiments are disclosed for receiving a target shape. The method may further include initializing a gradient mesh to a vector graphic having at least one node. The method may further include performing a constrained optimization of the vector graphic based on the target shape. The method may further include generating a stress metric based on a comparison of the constrained optimization and the target shape. The method may further include determining one or more unconstrained candidate vector graphics based on the stress metric. The method may further include selecting an improved vector graphic from the one or more unconstrained candidate vector graphics. The method may further include mapping the vector graphic to the improved vector graphic. The method may further include optimizing the improved vector graphic based on the target shape.

Continuous optimization of discrete parameters using a unified stress indicator

A method for manipulating a target image includes generating a query of the target image and keys and values of a first reference image. The method also includes generating matching costs by comparing the query of the target image with each key of the reference image and generating a set of weights from the matching costs. Further, the method includes generating a set of weighted values by applying each weight of the set of weights to a corresponding value of the values of the reference image and generating a weighted patch by adding each weighted value of the set of weighted values together. Additionally, the method includes generating a combined weighted patch by combining the weighted patch with additional weighted patches associated with additional queries of the target image and generating a manipulated image by applying the combined weighted patch to an image processing algorithm.

Image manipulation using deep learning techniques in a patch matching operation

Systems and methods described herein may relate to image transformation utilizing a plurality of deep neural networks. An example method includes receiving, at a mobile device, a plurality of image processing parameters. The method also includes causing an image sensor of the mobile device to capture an initial image and receiving, at a coefficient prediction neural network at the mobile device, an input image based on the initial image. The method further includes determining, using the coefficient prediction neural network, an image transformation model based on the input image and at least a portion of the plurality of image processing parameters. The method additionally includes receiving, at a rendering neural network at the mobile device, the initial image and the image transformation model. Yet further, the method includes generating, by the rendering neural network, a rendered image based on the initial image, according to the image transformation model.

Machine-learning based technique for fast image enhancement

Growing community of inventors

San Francisco, CA, United States of America

Michael Gharbi