US 8,554,011 B2Granted

Automatic exposure correction of images

AssigneeMicrosoft Corporation

InventorsLu Yuan, Jian Sun

Life of this patent

Estimating the term…

Loading patent…

Abstract

Techniques for automatic exposure correction of images are provided. In particular, the exposure of an input image may be improved by automatically modifying a non-linear function that characterizes the luminance of shadow, mid-tone, and highlight portions of the image. The input image may be segmented into a number of regions and each region is assigned a zone, where the zone indicates a specified range of luminance values. An initial zone assigned to a region of the image may be changed in order to reflect an optimal zone of the region. Based, in part, on the optimal zones for each region of the image, luminance modification parameters may be calculated and applied to the non-linear function in order to produce a modified version of the input image that improves the appearance of overexposed and/or underexposed regions of the input image.

Claims20

Dependent claims

An apparatus comprising:

one or more processors;
memory accessible to the one or more processors, the memory storing:
a segmentation and region analysis module executable by the one or more processors to determine a zone for each region of a plurality of regions of an image, the zone of each respective region indicating a specified range of luminance values associated with the respective region on a scale of luminance values;
an optimal zone assignment module executable by the one or more processors to modify the zone for at least one region of the plurality of regions such that the at least one region is associated with a respective modified zone;
a parameter calculation module executable by the one or more processors to calculate one or more luminance modification parameters for the image based, at least in part, on the zones for each respective region of the image and the respective modified zone of the at least one region;
and
a non-linear function adjustment module executable by the one or more processors to:
generate a non-linear function based, at least in part, on the one or more luminance modification parameters, the non-linear function specifying modifications of one or more luminance values of the image from an initial luminance value to a modified luminance value;
and
produce a modified version of the image according to the non-linear function by changing respective luminance values of pixels of the image from a respective initial luminance value to a respective modified luminance value.

7 more elements

A method comprising:

determining, by a computing device including one or more processors, a zone for each respective region of a plurality of regions of an image, the zone indicating a specified range of luminance values associated with the respective region on a scale of luminance values;
generating, by the computing device, a plurality of gamma corrected images based on the image, each of the plurality of gamma corrected images being generated using a different amount of gamma correction;
detecting, by the computing device, edges of a particular region of the image for each of the gamma corrected images and for the input image;
measuring, by the computing device, visible details of the particular region based, at least in part, on differences between edges of the particular region associated with a respective gamma-corrected image, edges of the particular region associated with at least one additional gamma-corrected image, edges of the particular region associated with the input image, or a combination thereof;
identifying, by the computing device, an additional region of the image that is adjacent to the particular region;
determining, by the computing device, a relative contrast between the particular region and the additional region based, at least in part, on luminance values of the particular region and luminance values of the additional region;
and
modifying the zone of the particular region based, at least in part, on the visible details of the particular region and the relative contrast.

5 more elements

Dependencies are inferred from the claim text.