Abstract: Systems and methods disclosed herein for providing people categorization within electronic images. One embodiment involves retrieving an input image from memory. The embodiment further involves determining one or more facial regions in the input image using a facial detection algorithm. The embodiment further involves identifying a number of features for each facial region. The embodiment further involves determining a subject face and a crowd face in the input image based on at least the number of features for the subject face being more than the number of features for the crowd face. The embodiment further involves displaying, on a display device, the subject face identified within the image.

Abstract: Embodiments relate to automatic perspective and horizon correction. Generally, a camera captures an image as an image file. Capture-time orientation data from one or more sensors is used to determine the camera's attitude with respect to a defined reference frame. The orientation data and/or attitude can be registered into metadata of the image file and used to generate axis lines representative of the camera's reference frame. A reference line such as a horizon can be automatically identified from detected line segments in the image that align with one of the axis lines within a predetermined angular threshold. The reference line can be used to generate a camera transformation from a starting orientation reflected by the camera attitude to a transformed orientation that aligns the reference line with the reference frame. The transformation can be applied to the image to automatically correct perspective distortion and/or horizon tilt in the image.

Abstract: Methods and systems are provided for performing automated capture of images based on emotion detection. In embodiments, a selection of an emotion class from among a set of emotion classes presented, via a graphical user interface, is received. The emotion class can indicate an emotion exhibited by a subject desired to be captured in an image. A set of images corresponding with a video is analyzed to identify at least one image in which a subject exhibits the emotion associated with the selected emotion class. The set of images can be analyzed using at least one neural network that classifies images in association with emotion exhibited in the images. Thereafter, the image can be presented in association with the selected emotion class.

Abstract: Techniques for removal of obstructions in a digital image scene are described, in which target and source digital images that exhibit parallax, one to another, are obtained that were captured together by an image capture device at a similar point in time using two different lenses of the image capture device. A foreground obstruction is identified based on displacement in apparent position of objects in the target and source digital images. The foreground obstruction is removed from the target digital image, such as by generating an obstruction mask that represents the location of the foreground obstruction and copying pixels from the source digital image to the target digital image based on the locations identified in the obstruction mask. The target digital image with the obstruction removed is output to a user interface or service provider system, for example.

Abstract: Techniques for converting mechanical markings on hardcopy textual content into digital annotations in a digital document file. In accordance with some embodiments, the techniques include identifying at least one block of text in a digital (scanned) image of a hardcopy document, and identifying at least one mechanical marking in the digital image of the hardcopy document. The mechanical marking, such as an underline, strike-through, highlight or circle, covers or lies adjacent to the corresponding block of text. An annotated digital document file is generated from the digital image of the hardcopy document. The annotated digital document file includes computer-executable instructions representing the original text of the hardcopy document and at least one annotation corresponding to the mechanical marking in the hardcopy document.

Abstract: An image editing application selectively edits a brushstroke in an image, based on a direction of the brushstroke. In some cases, the brushstroke is selectively edited based on a similarity between the direction of the brushstroke and a direction of an editing tool. Additionally or alternatively, directional data for each pixel of the brushstroke is compared to directional data for each position of the editing tool. Data structures capable of storing directional information for one or more of a pixel, a brushstroke, or a motion of an editing tool are disclosed.

Abstract: A user identifies an unwanted object in a source image. Related images are identified on the basis of timestamp and/or geolocation metadata. Matching masks are identified in the source image, wherein each of the matching masks is adjacent to the selection mask. Features in the selection and matching masks which also appear in one of the related images are identified. The related image having a maximum of features which are tracked to a source image matching mask, but also a minimum of features which are tracked to the source image selection mask, is identified as a best-match related image. By mapping the source image matching masks onto the best-match related image, a seed region can be located in the best-match related image. This seed region is used for filling in the source image. This allows the unwanted object to be replaced with a visually plausible background having a reasonable appearance.

Abstract: In implementations of image composition instruction based on reference image perspective, a template image of a desired object is captured. Reference images including a similar or same object as the desired object are obtained. Directions from a first location of where a template image was captured to a second location where a selected reference image was captured are obtained and exposed in a user interface to allow a user to move to the second location. At the second location, interactive instructions are generated based on a live video stream of the desired to move a camera capturing the video stream until a composition of the video stream is aligned with a composition of the reference image. The camera is configured with settings based on the reference image to capture a composed image having a same composition as the reference image.

Abstract: In implementations of image composition instruction based on reference image perspective, a template image of a desired object is captured. Reference images including a similar or same object as the desired object are obtained. Directions from a first location of where a template image was captured to a second location where a selected reference image was captured are obtained and exposed in a user interface to allow a user to move to the second location. At the second location, interactive instructions are generated based on a live video stream of the desired to move a camera capturing the video stream until a composition of the video stream is aligned with a composition of the reference image. The camera is configured with settings based on the reference image to capture a composed image having a same composition as the reference image.

Abstract: Methods, systems, and non-transitory computer readable storage media are disclosed for generating focused preview images that include the subjects of interest of digital images. For example, in one or more embodiments, the disclosed system utilizes a machine-learning model to generate a saliency map for a digital image to indicate one or more salient objects portrayed within the digital image. Additionally, in one or more embodiments, the system identifies a focus region based on focus information captured by a camera device at the time of capturing the digital image. Furthermore, the system can then utilize the saliency map and the focus region to generate a focused preview image. For instance, in one or more embodiments, the system crops the digital image based on an overlapping portion of the saliency map and the focus region to generate a focused preview image.

Abstract: Content-conforming stamp tool techniques are described. In one or more embodiments, a selection of an object in a digital image is received. An indication of a location in the digital image where the object is to be reproduced is also received. To reproduce and conform the object at the reproduction location, adjustments to a shape of the object are computed to conform a reproduction of the object to image content proximate the reproduction location. The adjustments are computed based on both the geometry of the image content at the source location and the geometry of the image content at the reproduction location. The adjustments are then applied to the shape of the object when it is reproduced at the reproduction location.

Abstract: Embodiments relate to automatic perspective and horizon correction. Generally, a camera captures an image as an image file. Capture-time orientation data from one or more sensors is used to determine the camera's attitude with respect to a defined reference frame. The orientation data and/or attitude can be registered into metadata of the image file and used to generate axis lines representative of the camera's reference frame. A reference line such as a horizon can be automatically identified from detected line segments in the image that align with one of the axis lines within a predetermined angular threshold. The reference line can be used to generate a camera transformation from a starting orientation reflected by the camera attitude to a transformed orientation that aligns the reference line with the reference frame. The transformation can be applied to the image to automatically correct perspective distortion and/or horizon tilt in the image.

Abstract: A digital medium environment includes at least one computing device. Systems and techniques are described herein for reassembling and repairing image sections (e.g., torn pieces of an image) by generating masks for the image sections that do not include content of the image sections, and matching image sections along contours (e.g., edges or sides of the image sections) based on shapes of the masks, features of content extracted from the image sections, or combinations thereof, depending on whether an ambiguity is determined. An ambiguity is determined when not all image sections included in the scans are matched by shape, or are redundantly matched. A composite image is reassembled from the image sections based on matching image sections. Furthermore, a composite image is repaired by adding material to the composite image from an additional image (e.g., an image other than the composite image and the image sections).

Abstract: Automated video correction techniques are disclosed. In some examples, an example method may include identifying features in each video frame of the multiple video frames in a video, and identifying one or more major scenes in the video based on a matching of the features in each video frame. The method may also include, for each identified major scene, identifying a key reference frame based on the features in each video frame, identifying one or more bad video frames based on a comparison with the key reference frame, and identifying one or more sequences of bad video frames based on the identified one or more bad video frames. The video may then be corrected by removing the identified one or more sequences of bad video frames from the video.

Abstract: An image editing application selectively edits a brushstroke in an image, based on a direction of the brushstroke. In some cases, the brushstroke is selectively edited based on a similarity between the direction of the brushstroke and a direction of an editing tool. Additionally or alternatively, directional data for each pixel of the brushstroke is compared to directional data for each position of the editing tool. Data structures capable of storing directional information for one or more of a pixel, a brushstroke, or a motion of an editing tool are disclosed.

Abstract: Automated video correction techniques are disclosed. In some examples, an example method may include identifying features in each video frame of the multiple video frames in a video, and identifying one or more major scenes in the video based on a matching of the features in each video frame. The method may also include, for each identified major scene, identifying a key reference frame based on the features in each video frame, identifying one or more bad video frames based on a comparison with the key reference frame, and identifying one or more sequences of bad video frames based on the identified one or more bad video frames. The video may then be corrected by removing the identified one or more sequences of bad video frames from the video.

Abstract: The present disclosure is directed towards systems and methods for modifying a digital image. For example, systems and methods described herein involve identifying a target portion of a digital image to remove from the digital image. The systems and methods further involve identifying geometric features intersecting or surrounding the identified target portion. The systems and methods further involve analyzing and identifying a source portion having geometric properties that correspond to the identified geometric features intersecting the target portion. Further, the systems and methods involve removing the target portion and replacing the target portion with the source portion by aligning the geometric properties to blend together with the background surrounding the removed target portion. In this way, the output image includes a replacement portion in place of the removed target portion that blends together with the background of the output image.

Abstract: Methods, systems, and non-transitory computer readable storage media are disclosed for generating focused preview images that include the subjects of interest of digital images. For example, in one or more embodiments, the disclosed system utilizes a machine-learning model to generate a saliency map for a digital image to indicate one or more salient objects portrayed within the digital image. Additionally, in one or more embodiments, the system identifies a focus region based on focus information captured by a camera device at the time of capturing the digital image. Furthermore, the system can then utilize the saliency map and the focus region to generate a focused preview image. For instance, in one or more embodiments, the system crops the digital image based on an overlapping portion of the saliency map and the focus region to generate a focused preview image.

Abstract: A digital medium environment includes at least one computing device. Systems and techniques are described herein for reassembling and repairing image sections (e.g., torn pieces of an image) by generating masks for the image sections that do not include content of the image sections, and matching image sections along contours (e.g., edges or sides of the image sections) based on shapes of the masks, features of content extracted from the image sections, or combinations thereof, depending on whether an ambiguity is determined. An ambiguity is determined when not all image sections included in the scans are matched by shape, or are redundantly matched. A composite image is reassembled from the image sections based on matching image sections. Furthermore, a composite image is repaired by adding material to the composite image from an additional image (e.g., an image other than the composite image and the image sections).

Abstract: A context aware clipping mask is described. Context aware clipping masks may provide an automated approach for creating clipping sets that generate aesthetically pleasing photo texts. In at least some implementations, an input image is analyzed to generate an image saliency map and an image gradient map. A plurality of image rects of varying sizes and aspect ratios are generated on a grid over the input image. A score is assigned to each image rect, incorporating a visual composition, a boundary simplicity based on the image gradient map, and a content preservation based on the image saliency map. An image rect with the highest score is selected and utilized to create a photo text from the input image.