Abstract: A waveguide image combiner is used to transmit a monochrome or full-color image in an augmented reality display. The combiner uses multiple stacked substrates and multiple pairs of incoupling and outcouping VHOEs to expand a first FOV and an image expander to expand the second or perpendicular FOV. This suitably provides an expanded FOV that offers a diagonal FOV?50°, a horizontal FOV?40 and a vertical FOV?25°. The combiner also delivers a large horizontal eye box up to 20 mm and a vertical eye box of 10 mm while maintaining high light efficiency of the real scene (e.g. >80%). The system is able to use a light engine based on broadband (10 nm????40 nm) LEDs and maintain a large horizontal field of view and high transmission of the real imagery. The approach resolves issues with current embodiments including astigmatism, image overlap, color balance, and small light engine pupils leading to reduced eye boxes.

Abstract: Various embodiments of a framework which allow, as an alternative to resource-taxing decompression, efficient computation of feature maps using a compressed content data subset, such as video, by exploiting the motion information, such as a motion vector, present in the compressed video. This framework allows frame-specific object recognition and action detection algorithms to be applied to compressed video and other media files by executing only on I-frames in a Group of Pictures and linearly interpolating the results. Training and machine learning increases recognition accuracy. Yielding significant computational gains, this approach accelerates frame-wise feature extraction I-frame/P-frame/P-frame videos as well as I-frame/P-frame/B-frame videos. The present techniques may also be used for segmentation to identify and label respective regions for objects in a video.

Abstract: Images of various views of objects can be captured. An object mesh structure can be created based at least in part on the object images. The object mesh structure represents the three-dimensional shape of the object and includes a mesh with mesh elements. The mesh elements are assigned views first from a subset of views to texture large contiguous portions of the object from relatively few views. Portions that are not textured from the subset views are textured using the full set of views, such that all mesh elements are assigned views. The views first assigned from the subset of views and the views then assigned from the full plurality of views can be packaged into a texture atlas. These texture atlas views can be packaged with mapping data to map the texture atlas views to their corresponding mesh elements. The texture atlas and the object mesh structure can be sent to a client device to render a representation of the object.

Abstract: Customizable intrusion zones for audio/video (A/V) recording and communication devices in accordance with various embodiments of the present disclosure are provided. In one embodiment, a method for an A/V recording and communication device is provided, the method comprising displaying, on a display of the computing device, a user interface for creating and/or customizing at least one intrusion zone, wherein the at least one intrusion zone comprises at least one motion zone within a field of view of the A/V recording and communication device coupled with at least one conditional setting of the at least one motion zone, determining whether an input has been received to establish a new conditional setting, or to change a previous conditional setting, determining whether an input has been received to save the new conditional setting or to save the changed conditional setting, and saving the new conditional setting or the changed conditional setting.

Abstract: System and methods are provided that generate a three-dimensional model from a physical space. While a user is scanning and/or recording the physical space with a user computing device, user speech describing the physical space is recorded. A transcript is generated from the audio captured during the scan and/or image recording of the physical space. Keywords from the transcript are used to improve computer-vision object identification, which is incorporated in the three-dimensional model.

Abstract: Large scale instance recognition is provided that can take advantage of channel-wise pooling. A received query image is processed to extract a set of features that can be used to generate a set of region proposals. The proposed regions of image data are processed using a trained classifier to classify the regions as object or non-object regions. Extracted features for the object regions are processed using feature correlation against extracted features for a set of object images, each representing a classified object. Matching tensors generated from the comparison are processed using a spatial verification network to determine match scores for the various object images with respect to a specific object region. The match scores are used to determine which objects, or types of objects, are represented in the query image. Information or content associated with the matching objects can be provided as part of a response.

Abstract: Techniques for placing and manipulating multiple three-dimensional (3D) models using mobile augmented reality (AR) are described. One technique includes receiving a first request to initialize an AR simulation of a first product for sale within a physical environment. In response to the first request, a first 3D model of the first product for sale is rendered onto the screen. After rendering the first 3D model, a second request to visualize a second product for sale within the physical environment is received during the AR simulation of the first product for sale. In response to the second request, a second 3D model of the second product for sale is rendered onto the screen with the first 3D model.

Abstract: Various embodiments of systems and methods for mapping review articles to locations in a marketplace catalog. Various embodiments include training a system using product reviews and images from the marketplace as training data. Product reviews and images can be provided to autoencoders to generate respective short codes. The product review short codes can be combined with the product image short codes to generate review short codes. These review short codes can be provided as training data for an attribute assignment network while associated attributes (e.g., attributes associated with related products) can be target outputs for the attribute assignment network. After the system is sufficiently trained, a review article can be obtained from the publishing platform. This review article can be provided to the trained system to receive predicted attribute assignments. These predicted attribute assignments can be effective to map the review article to locations in a marketplace catalog.

Abstract: Various approaches discussed herein enable browsing groups of visually similar items to an item of interest, wherein the item of interest may be identified in a query image, for example. One or more visual attributes associated with the item of interest are identified, and the visually similar items matching at least one of the visual attributes are grouped together, wherein the group is ranked according to the visually similar items' overall visual similarity to the item of interest, for example by using a visual similarity score and/or metric.

Abstract: Search results are filtered using one or more rankers that evaluate a searcher intention to select items from a subset of search results. A subset of items may be provided that are responsive to a user search query. This subset of items may have one or more related properties and may be logically grouped together. A database of rankers may be evaluated and applied to the subset of items to determine items that are both relevant to the user search query and also correspond to a user intention for the search. As a result, the ranker may select and filter out certain items having one or more features that do not correspond to the user intention.

Abstract: Users can view images or renderings of items placed (virtually) within a physical space. For example, a rendering of an item can be placed within a live camera view of the physical space. A snapshot of the physical space can be captured and the snapshot can be customized, shared, etc. The renderings can be represented as two-dimensional images, e.g., virtual stickers or three-dimensional models of the items. Users can have the ability to view different renderings, move those items around, and develop views of the physical space that may be desirable. The renderings can link to products offered through an electronic marketplace and those products can be consumed. Further, collaborative design is enabled through modeling the physical space and enabling users to view and move around the renderings in a virtual view of the physical space.

Abstract: Various approaches provide for visual similarity-based search techniques that allow a user to provide two-dimensional image data (e.g., still images or video) about an item and search for items having related three-dimensional features. In order to create an electronic catalog of items that is searchable by three-dimensional features, the features are extracted from image data for each item. The three-dimensional features can be extracted using a trained model, such as a trained neural network or other machine learning-based approach. Thus, three-dimensional features of items in the electronic catalog can be determined and used to select or rank the items in response to a visual search query based on determined three-dimensional features of the query item. For example, three-dimensional features of the query item can be used to query the electronic catalog of items, in which items having visual attributes of similar three-dimensional features are selected and returned as search results.

Abstract: Various embodiments of systems and methods allow a distributed database to be maintained to store records of activities that may promote purchasing a product and determine whether such promotions should be rewarded. Using anonymous identifiers and a permission-based system, identities of users, products, content, content distribution systems, and marketplaces can be made available only to parties with rights to know such identities. Furthermore, systems can be put in place so that activities with a certain product, content item, etc. can be monitored without ever disclosing the identity of the product, content item, etc. In some embodiments, the database can be an immutable, append-only, distributed database such as a blockchain ledger.

Abstract: Systems and methods are disclosed for refining the accuracy of network searches by supplementing existing keywords and key phrases in an e-commerce catalog or other database with aggregated and analyzed additional, external data. The internet or another network can be crawled for identifiers which point to entries in the catalog or other database, and, subject to third-party use restrictions, data and metadata can be extracted to enrich the existing keywords and key phrases. The extracted external content may be processed by machine learning techniques in order to find similar entries in the original catalog or database. Categorizing and indexing the entries further improves search recall, including clustering via processing word embeddings.

Abstract: Supplemental content can be determined through a server-side process, wherein a primary content provider can obtain the supplemental content from a supplemental content provider before any of the content is sent to the client device initially requesting the primary content. A primary content provider submits a request that can include an authorization token for the primary content provider, a publisher token including a proxy identifier for the user, application, or client device, and other relevant information. The supplemental content provider can verify the information and select targeted content based on the information. The supplemental content can then be provided to the primary content provider for incorporation with the primary content. The primary content provider can adjust the layout or selection before sending to the client device for display.

Abstract: A method for network configuration sharing includes receiving a home network configuration associated with a user, identifying an away network, and modifying the configuration of the away network based on the configuration of the home network.

Abstract: A system and method are provided that present users with a platform to complete and self-file customs entry forms for goods transported across an international border, wherein the system and method enable the user to resolve a flagged exception from a customs agency through an automated computerized response or by using customs-broker resources via the digital network. The system further enables a user to validate submitted and cleared custom entry forms by setting parameters to dictate a level of auditing conducted by customs-broker resources. As such, users can readily complete and file the necessary customs entry forms while maintaining customs-broker resources on the back-end for selective auditing and resolving issues with a customs agency, thereby lowering associated overhead costs through a reduced dependency on customs-broker resources.

Abstract: Various embodiments provide for determining problematic components of an application that may cause user experience issues when running the application, and preemptively turning off or disabling the problematic components so that the application is rendered without the problematic component. When a component is rendered, various runtime data is produced. Such data may be logged and analyzed, and used to generate performance metrics of the component. Such performance metrics can be used to determine whether a component should be disabled. When a component is disabled, it become unavailable and will be excluded from the application when the application is rendered and/or served to a user. Thus, when an application is rendered or served, only available or enabled components are included.

Abstract: System and methods are provided that can address cold-start problems in database keyword searches. The search system generates machine-learned values for new items based on historical signals for already existing items. These initial values are generated at the time of new item's inclusion in the search index. The values are used as input in a ranking model to rank search results for a user query. The initial values for the new items predict user engagement with the new items based on historical data for existing items and increase the visibility of new items to accumulate user interaction data for the new items.

Abstract: Techniques for improving the visualization of content with a mixed reality (MR) device are described. One techniques includes monitoring a gaze of a user of a MR device through one or more displays of the MR device. A convergence distance of the gaze of the user is determined based on the monitoring. A selection between a first projection mode and a second projection to display content on the one or more displays is made based on whether the convergence distance exceeds a threshold distance from the user. Content is displayed on the one or more displays, based on the selection.