Examples disclosed herein may involve (i) obtaining a first layer of map data associated with sensor data capturing a geographical area, the first layer of map data comprising an aggregated overhead-view image of the geographical area, where the aggregated overhead-view image is generated from aggregated pixel values from a plurality of images associated with the geographical area, (ii) obtaining a second layer of map data, the second layer of map data comprising label data for the geographical area derived from the aggregated overhead-view image of the geographical area, and (iii) causing the first layer of map data and the second layer of map data to be presented to a user for curation of the label data.

A computer-implemented method includes receiving one or more images from one or more sensors, creating one or more image patches utilizing the one or more images, creating one or more latent representations from the one or more image patches via a neural network, outputting, to a concept extractor network, the one or more latent representations utilizing the one or more image patches, defining one or more scores associated with the one or more latent representations, and outputting one or more scores associated with the one or more image patches utilizing at least the concept extractor network.

A method of remote facilities monitoring for the detection of contamination, leaks or failures in petrochemical and related facilities. Remote cameras are used to capture images of equipment to be inspected, which are transmitted to a central server. A software component on the central server applies artificial intelligence and image processing methods to detect the presence of anomalies in the images for display to an operator and potential subsequent dispatch of in person follow-up. Parallel data streams from sensors at the remote sites can be used by the software on the server to enhance the level of confidence in anomaly detection. The rl server software uses an adaptive object-detection function to analyze image data and learns over time to provide enhanced detection of interest regions and failure conditions.

System and method for crowdsourcing a video summary for creating an enhanced video summary are disclosed. The method includes receiving videos, analysing the videos, creating the video summary of the videos using a building block model, storing the video summary in a video library database, crowdsourcing the video summary to at least one of the plurality of users, enabling the at least one of the plurality of users to review the video summary and identify at least one new characteristic, enabling the at least one of the plurality of users to share the at least one new characteristic on the platform, comparing at least one existing characteristic of the building block model with the corresponding new characteristic, reconciling the video summary along with at least one inserted new characteristic, creating a new building block model, editing the video summary for creating the enhanced video summary.

A user's profile picture is updated from a live video stream. A profile updater analyzes the still images that make up the live video stream and identifies one or more target characteristics of a subject in the still images. Using the target characteristics, one or more still images are selected for use as the updated profile picture. The user provides feedback regarding the selected images, which is used to refine the still image selection.

Data-dependent node-to-node knowledge sharing to increase the interpretability of the activation pattern of one or more nodes in a neural network, is implemented by a set of knowledge sharing links. Each link may comprise a knowledge providing node or other source P and a knowledge receiving node R. A knowledge sharing link can impose a node-specific regularization on the knowledge receiving node R to help guide the knowledge receiving node R to have an activation pattern that is more easily interpreted. The specification and training of the knowledge sharing links may be controlled by a cooperative human-AI learning supervisor system in which a human and an artificial intelligence system work cooperatively to improve the interpretability and performance of the client system.

An image processing apparatus includes a character recognition processing unit configured to execute character recognition processing on the image data, an acquisition unit configured to acquire one or more character string blocks included in the image data, from the image data, a selection unit configured to select a character string block to be used for setting of a file name, from among the one or more character string blocks acquired by the acquisition unit, and a setting unit configured to set the file name of image data by using a character recognition result of the character recognition processing unit for the character string block selected by the selection unit.

A method identifies an article by associating an image of the article with stored article data. The method includes: A) creating the image of the article; B) extracting image features from the image; and C) extracting article features from the stored article data. The method step C) are configured to be carried out before, during, or after method steps A) and B). The following method steps are then carried out: D) comparing the image features to the article features; E) output of association information of the image with the stored article data.

Methods and systems for transforming at least a portion of a physical document into digital data. One method includes obtaining a first plurality of data items automatically extracted from a first physical document and a validated value for a data field. The method also includes automatically identifying a first linked data item included in the first plurality of data items that is linked to the validated value and setting a physical position included in a rule to the physical position of the first linked data item. In addition, the method includes obtaining a second plurality of data items automatically extracted from a second physical document and automatically identifying a candidate data item included in the second plurality of data items based on the rule. Furthermore, the method includes automatically populating a value for the data field for the second physical document based on the candidate data item.

Techniques are described herein that enable advanced gaming and virtual reality control using radar. These techniques enable small motions and displacements to be tracked, even in the millimeter or submillimeter scale, for user control actions even when those actions are optically occluded or obscured.