Image adjustment
AssigneeAdobe Systems Incorporated
InventorsSylvain P. Paris, Frederic P. Durand, Vladimir L. Bychkovsky, Eric Chan
Life of this patent
Estimating the term…
Abstract
Techniques are disclosed relating to automatically adjusting images. In one embodiment, an image may be automatically adjusted based on a regression model trained with a database of raw and adjusted images. In one embodiment, an image may be automatically adjusted based on a model trained by both a database of raw and adjusted images and a small set of images adjusted by a different user. In one embodiment, an image may be automatically adjusted based on a model trained by a database of raw and adjusted images and predicted differences between a user's adjustment to a small set of images and a predicted adjustment based on the database of raw and adjusted images.
Claims20
A computer-implemented method comprising;
- receiving, by a computing device, one or more adjustments made by a user to one or more parameters of each raw image in a subset of a plurality of raw images, each raw image in the subset of the plurality of raw images corresponding to an image pair that includes the raw image and an adjusted version of the raw image having one or more adjusted parameters in comparison to one or more unadjusted parameters of the raw image;
- computing, by the computing device, a transferred adjustment that is configured to model the one or more adjustments made by the user and which is usable for subsequent automatic adjustment of new images;
- computing an adjustment offset that defines a difference between the transferred adjustment and a plurality of curves that describe the one or more adjusted parameters corresponding to each image pair;
- receiving, by the computing device, a new image;
- and
- automatically globally adjusting, by the computing device, the new image by at least applying the transferred adjustment and the adjustment offset to the new image to create an adjusted new image that is based on the one or more modeled adjustments.
A non-transitory computer readable storage device storing program instructions that are executable by a computer to cause the computer to perform a method comprising:
- applying a predictive adjustment to a new image based on a reference predicted adjustment, the reference predicted adjustment being based on a plurality of curves that each relate one or more unadjusted parameters of a raw image captured from an image capturing device to one or more adjusted parameters of an adjusted image of an image pair;
- computing a predictive correction that defines a difference between the reference predictive adjustment and adjustment preferences of a user, the adjustment preferences of the user being based on one or more adjustments made by the user to one or more parameters of each raw image in a set of raw images;
- and
- applying the predictive correction to the new image, subsequent to the predictive adjustment being applied, to generate an adjusted new image.
A system comprising:
- one or more processors and at least one memory that maintains instructions that are executable by the one or more processors to cause operations to be performed comprising:
- receiving one or more adjustments to each raw image of a subset of raw images of a plurality of image pairs, each said image pair comprising a raw image and an adjusted version of the raw image having one or more adjusted parameters in comparison to one or more corresponding unadjusted parameters of the raw image,
- computing a transferred adjustment by correlating a trained regression algorithm to the one or more adjustments of the subset of raw images;
- computing an adjustment offset that defines a difference between the transferred adjustment and one or more adjustment preferences of a user;
- and
- automatically globally adjusting one or more parameters of a new image by at least applying the transferred adjustment and the adjustment offset to the new image to create an adjusted new image.
Dependencies are inferred from the claim text.