Abstract: Systems, devices, media, and methods are presented for graphical icon identification within an image or video stream. The systems and methods receive an image including a graphical icon. The systems and methods identify a set of proposed regions of the image, at least one proposed region of the set of proposed regions containing the graphical icon and extract a set of semantic features for each proposed region of the set of proposed regions. Based on the set of semantic features of the set of proposed regions, the systems and methods identify a set of proposed icons corresponding to the graphical icon included in the image and determine a match between the graphical icon and at least one proposed icon of the set of proposed icons.

Abstract: Systems, devices, media, and methods are presented for graphical icon identification within an image or video stream. The systems and methods receive an image including a graphical icon. The systems and methods identify a set of proposed regions of the image, at least one proposed region of the set of proposed regions containing the graphical icon and extract a set of semantic features for each proposed region of the set of proposed regions. Based on the set of semantic features of the set of proposed regions, the systems and methods identify a set of proposed icons corresponding to the graphical icon included in the image and determine a match between the graphical icon and at least one proposed icon of the set of proposed icons.

Abstract: A machine learning system can generate an image mask (e.g., a pixel mask) comprising pixel assignments for pixels. The pixels can he assigned to classes, including, for example, face, clothes, body skin, or hair. The machine learning system can be implemented. using a convolutional neural network that is configured to execute efficiently on computing devices having limited resources, such as mobile phones. The pixel mask can be used to more accurately display video effects interacting with a user or subject depicted in the image.

Abstract: Symbol prediction can be implemented using a multi-task system trained for different tasks. The tasks may include a single symbol prediction, symbol category prediction, and symbol subcategory prediction. Categories of symbols can be generated by clustering sets of training data using a clustering scheme.

Abstract: Because not all companies register or obtain barcodes (or other encodings or identifiers) from a single, consistent source, and because different companies can be associated with the same barcode at different times, there can be multiple products associated with a specific barcode. If a user requests information associated with a specific barcode, information associated with the user or request can be used to attempt to determine the associated product if more than one matching product is uncovered. In some embodiments, a location of a user device at approximately a time of the request can be used to attempt to determine the appropriate product, such as if the user is in a store that only offers one of the potential products. Other information can be used as well, such as past locations, purchase history, viewing patterns, and the like.

Abstract: Various embodiments provide methods and systems for users and business owners to share content and/or links to visual elements of a place at a physical location, and, in response to a user device pointing at a tagged place, causing the content and/or links to the visual elements of the place to be presented on the user device. In some embodiments, content and links are tied to specific objects at a place based at least in part upon one of Global Positioning System (GPS) locations, Inertial Measurement Unit (IMU) orientations, compass data, or one or more visual matching algorithms. Once the content and links are attached to the specific objects of the place, they can be discovered by a user with a portable device pointing at the specific objects in the real world.

Abstract: Various embodiments provide methods and systems for users and business owners to share content and/or links to visual elements of a place at a physical location, and, in response to a user device pointing at a tagged place, causing the content and/or links to the visual elements of the place to be presented on the user device. In some embodiments, content and links are tied to specific objects at a place based at least in part upon one of Global Positioning System (GPS) locations, Inertial Measurement Unit (IMU) orientations, compass data, or one or more visual matching algorithms. Once the content and links are attached to the specific objects of the place, they can be discovered by a user with a portable device pointing at the specific objects in the real world.

Abstract: Symbol prediction can be implemented using a multi-task system trained for different tasks. The tasks may include a single symbol prediction, symbol category prediction, and symbol subcategory prediction. Categories of symbols can be generated by clustering sets of training data using a clustering scheme.

Abstract: A machine learning system can generate an image mask (e.g., a pixel mask) comprising pixel assignments for pixels. The pixels can be assigned to classes, including, for example, face, clothes, body skin, or hair. The machine learning system can be implemented using a convolutional neural network that is configured to execute efficiently on computing devices having limited resources, such as mobile phones. The pixel mask can be used to more accurately display video effects interacting with a user or subject depicted in the image.

Abstract: Various embodiments relating to predictive analytics from visual data are described. For example, a system associated with a matching service can obtain visual data that includes content related to a marketing campaign associated with a particular product or service. The system can determine image information that contains one or more objects relating to a logo for the particular brand or product. The system can process the content item using one or more machine vision algorithms. The system can determine usage data associated with the particular product or service based at least in part on the one or more feature descriptors. The system can aggregate the usage data with other usage data relating to different users across different time periods to compile a database of insight information with respect to the particular product or service. The system can determine one or more usage patterns relating to the particular product or service.

Abstract: Systems and methods for displaying an image of a virtual object in an environment are described. A computing device is used to capture an image of a real environment including a marker. One or more virtual objects which do not exist in the real environment are displayed in the image based at least on the marker. The distance and orientation of the marker may be taken into account to properly size and place the virtual object in the image. Further, virtual lighting may be added to an image to indicate to a user how the virtual object would appear with the virtual lighting.

Abstract: Embodiments of the present disclosure relate generally to systems for presenting content suggestions within a messaging interface. A content suggestion system performs operations that include: receiving inputs that include a text string into a messaging interface; parsing a keyword from the text string of the input in response to receiving the inputs; retrieving message content that includes at least a media item based on the keyword; and presenting a chat suggestion that includes the message content within the messaging interface, according to certain example embodiments.

Abstract: Approaches enable images submitted by users, owner, and/or authorized person of a point of interest (e.g., a place, a scene, an object, etc.) to be used as a fiducial to assist recognition and tracking of the point of interest in an augmented reality environment. Multiple images (e.g., crowd-sourced images) of a point of interest taken from different points of view can be dynamically used. For example, as a user with a user device moves through a point of interest, a different image can be chosen from a set of stored candidate images of the point of interest based at least upon GPS locations, IMU orientations, or compass data of the user device. In this way, instead of relying on artificial fiducial images for various detection and tracking approaches, approaches enable images submitted by users and/or an owner or other authorized person of a point of interest to be used as fiducials to assist recognition and tracking of the point of interest.

Abstract: A content suggestion system to generate and cause display of a set of chat suggestions based on messages received at a client device. The content suggestions system is configured to display messages that include message content at a client device, and identify content selected by a user of the client device to be included in a response to the messages received at the client device. The content suggestion system tracks and stores a number of times in which a particular pair of content appear in succession in a chat context, and calculates a ranking of the content among a set of available content. When subsequent messages that include the content of the content pair are displayed at the client device, the content suggestion system retrieves and presents a set of content as suggestions, based on the corresponding ranks.

Abstract: A machine learning system can generate an image mask (e.g., a pixel mask) comprising pixel assignments for pixels. The pixels can be assigned to classes, including, for example, face, clothes, body skin, or hair. The machine learning system can be implemented using a convolutional neural network that is configured to execute efficiently on computing devices having limited resources, such as mobile phones. The pixel mask can be used to more accurately display video effects interacting with a user or subject depicted in the image.

Abstract: Various embodiments provide methods and systems for users and business owners to share content and/or links to visual elements of a place at a physical location, and, in response to a user device pointing at a tagged place, causing the content and/or links to the visual elements of the place to be presented on the user device. In some embodiments, content and links are tied to specific objects at a place based at least in part upon one of Global Positioning System (GPS) locations, Inertial Measurement Unit (IMU) orientations, compass data, or one or more visual matching algorithms. Once the content and links are attached to the specific objects of the place, they can be discovered by a user with a portable device pointing at the specific objects in the real world.

Abstract: Various approaches discussed herein enable augmenting an image of a person wearing an article of clothing to provide an augmented reality (i.e., virtual) article of clothing to a viewer in order to overlay, replace, and/or augment an actual article of clothing being worn by the user. The augmented reality article of clothing is positioned on the user's real-world piece of clothing according to various markings and/or data that allow the augmented reality article of clothing to appear as if it were being worn by the user, even during movement of the user.

Abstract: Various embodiments provide methods and systems for users and business owners to share content and/or links to visual elements of a place at a physical location, and, in response to a user device pointing at a tagged place, causing the content and/or links to the visual elements of the place to be presented on the user device. In some embodiments, content and links are tied to specific objects at a place based at least in part upon one of Global Positioning System (GPS) locations, Inertial Measurement Unit (IMU) orientations, compass data, or one or more visual matching algorithms. Once the content and links are attached to the specific objects of the place, they can be discovered by a user with a portable device pointing at the specific objects in the real world.

Abstract: Approaches enable images submitted by users, owner, and/or authorized person of a point of interest (e.g., a place, a scene, an object, etc.) to be used as a fiducial to assist recognition and tracking of the point of interest in an augmented reality environment. Multiple images (e.g., crowd-sourced images) of a point of interest taken from different points of view can be dynamically used. For example, as a user with a user device moves through a point of interest, a different image can be chosen from a set of stored candidate images of the point of interest based at least upon GPS locations, IMU orientations, or compass data of the user device. In this way, instead of relying on artificial fiducial images for various detection and tracking approaches, approaches enable images submitted by users and/or an owner or other authorized person of a point of interest to be used as fiducials to assist recognition and tracking of the point of interest.

Abstract: Various embodiments provide methods and systems for detecting one or more segments of an image that are related to a particular object in the image (e.g., a logo or trademark) and extracting at least one feature point, each of which is represented by one feature point descriptor, based at least upon a contour curvature of the one or more segments. The at least one feature point descriptor can be converted into one or more codewords to generate a codeword database. A discriminative codebook can then be generated based upon the codeword database and utilized to detect objects and/or features in a query image.