Google Inc.

Sevket Derin Babacan

Aravind Krishnaswamy

Krzysztof Kulewski

Stephen J DiVerdi

Tinglin Liu

Bo-Jhang Ho

Justin Lewis

Gavin James

Daniel Pettigrew

Jason Chang

Grayson Squier Lang

Ronald Frank Wotzlaw

Timothy Sepkoski St Clair

John Thomas Nack

Juan Carlos Miguel Anorga

Stephen Joseph Diverdi

Jordi Miguel Verdu Orts

Hao Xu

Implementations relate to visualizations including images based on image content. In some implementations, a computer-implemented method includes obtaining a set of images, determining one or more pixel characteristics of the set of images, and determining one or more faces depicted in the plurality of images based on one or more pixel characteristics. The method selects a group of images of the set of images, where each image in the group of images depicts a different group of faces than depicted in the other images in the set of images. The method generates a visualization including the group of images, and provides the visualization to a user device in response to a user request to cause the group of images to be displayed by the user device.

Visualization of image themes based on image content

Implementations generally relate to providing regions of interest in an image. In some implementations, a method includes receiving an image. The method further includes determining one or more image features. The method further includes grouping the one or more image features into one or more regions of interest.

Providing regions of interest in an image

Implementations relate to computer-assisted cropping of an image. A computer-executed method includes receiving user input indicative of a change in size of a crop window from a first size to a second size, where the crop window is displayed over an image in an image editing user interface. The method detects that the crop window of the second size has an aspect ratio within a threshold range of a predetermined aspect ratio. In response to the detection, the method resizes the crop window to a third size, where the crop window of the third size has a resized aspect ratio substantially the same as the predetermined aspect ratio.

Computer-assisted image cropping

Methods and systems for modifying an image by applying an effect to an image are described. The effects include a pop effect, a light adjustment, or a color adjustment to an image. The methods and systems include providing a user slider for applying an effect to the image. The methods and systems further include determining a first portion of the image including a face and creating a protection mask to protect the face in the first portion during image modification. The protection mask may include an enhancement threshold for modifying the first portion of the image. The modification of the image may include modifying the second portion of the image differently than the first portion of the image. A method for enforcing different resolutions of a same input image to produce similar visual results is also described.

User sliders for simplified adjustment of images

Implementations relate to providing visual effects for images. In some implementations, a method includes detecting one or more objects in an image. The method identifies one or more important objects of the objects, where the important objects are determined to have an importance measurement satisfying a predetermined threshold indicating their importance to a viewer of the image. The method determines an application of a visual image effect to the image based on the important objects.

Providing visual effects for images

Implementations relate to visualizing and measuring impact of image modifications. In some implementations, a method to measure and indicate impact of image modification includes applying an edit operation to a first image, including modifying one or more pixels of the first image to provide a modified image. The method determines an impact score associated with the edit operation and indicative of a degree of visual impact of the edit operation to the first image. The method provides, based on the impact score, the modified image in a visualization of image modification for the first image, and provides the visualization for display by a display device.

Measuring and visualizing impact of image modifications

Implementations relate to adjusting images using a texture mask. In some implementations, a method includes detecting one or more texture regions having detected texture in an image, and generating a mask from the image based on the detected texture regions. The detected texture regions are distinguished in the mask from other regions of the image that do not have detected texture. The method applies one or more adjustment operations to the image in amounts based on values of the mask.

Image adjustment using texture mask

Implementations generally relate to enhancing a video. In some implementations, a method includes classifying one or more objects in one or more frames of the video. The method further includes determining one or more filter parameters of one or more filters based on the classifying of the one or more objects. The method further includes smoothing one or more of the determined filter parameters based on the classifying of the one or more objects. The method further includes applying one or more of the filters with corresponding smoothed filter parameters to one or more frames of the video.

Temporally smooth video enhancement

Aspects of the subject technology relate to automatically and selectively applying a fill light filter to an image. A process includes determining an edge-preserved, smoothed version of the image, and determining a grayscale version of the image. The process also includes comparing each pixel of the edge-preserved, smoothed version of the image to each corresponding pixel of the grayscale version of the image. The process also includes applying the fill light filter to the image based on the comparison. The fill light filter is automatically adjusted based on identifying regions in the image. Selectively applying the fill light filter can reduce artifacts and noise from forming or being amplified as a result of the fill light filter.

Automated selective application of fill light filter

Systems, methods and computer readable media for determination of monochromatic images are described. Some implementations can include a method. The method can include converting an image in a first colorspace to a first converted image in a second colorspace. The method can also include generating a hue histogram based on hue values in the converted image and determining a hue dispersion measure based on the hue histogram and a number of pixels in the image. The method can further include determining a hue clustering value based on the hue histogram and converting the image in the first colorspace into a second converted image in a third colorspace. The method can also include determining a variance measure based on the second converted image, and comparing the hue dispersion measure, the hue clustering value and the variance measure to a first threshold value, a second threshold value and a third threshold value, respectively.

Monochromatic image determination

Implementations relate to estimating noise in images using a polynomial relationship for pixel values of image features. In some implementations, a computer-implemented method to estimate noise in an image includes determining a plurality of patches of pixels in the image. For each patch of pixels, the method determines feature pixels in the patch that are included in a particular image feature at least partially depicted in the patch. The method determines an error estimate for each patch of pixels, where each error estimate is based on an amount by which pixel values of the feature pixels in the patch of pixels are different from an estimated polynomial relationship between the feature pixels in the patch of pixels. One of the error estimates is selected as a noise level estimate for the image.

Noise estimation for images using polynomial relationship for pixel values of image features

Implementations relate to detecting and modifying facial features of persons in images. In some implementations, a method includes receiving one or more general color models of color distribution for a facial feature of persons depicted in training images. The method obtains an input image, and determines a feature mask associated with the facial feature for one or more faces in the input image. Determining the mask includes estimating one or more local color models for each of the faces in the input image based on the general color models, and iteratively refining the estimated local color models based on the general color models. The refined local color models are used in the determination of the feature mask. The method applies a modification to the facial feature of faces in the input image using the feature mask.

Detecting and modifying facial features of persons in images

In some implementations, a method includes identifying one or more face regions of an image, the face regions including pixels that depict at least a portion of one or more faces of persons. The face regions are identified based on identifying facial landmarks of the faces. The method determines an associated face mask for each of the faces based on the face regions, where each face mask indicates which pixels in the image depict the corresponding face. Face pixels can be selected for processing by applying the face masks, and image pixels outside the faces can be selected by inversely applying the face masks. The selected pixels can be provided to a processing operation for adjustment of the selected pixels.

Image masks for face-related selection and processing in images

Implementations relate to estimating noise in images. In some implementations, a method includes extracting a plurality of sample blocks of pixels from a received image, where each sample block includes a subset of pixels of the image. One or more of the sample blocks are examined for texture content based on whether the sample blocks include one or more edges based on a predetermined threshold. At least one sample block determined to include texture content is removed. The method determines one or more average color variances based on the remaining sample blocks that have not been removed, where noise estimations for the image are based on the average color variances.

Growing community of inventors

San Francisco, CA, United States of America

Sevket Derin Babacan