Abstract: A method implemented at a computing system for detecting anomalies in video frames. The method includes a temporal dependency machine-learning (ML) model determining a set of temporal features that encapsulate temporal dependencies across received video frames at the computing system; A semantic injection ML model injects semantic information into the set of temporal features to thereby generate a set of enhanced temporal features. The semantic information defines contextual information about visual content of the received video frames.

Abstract: One example method is a method for a virtual shopping experience. The method includes connecting an app on a mobile device of a customer to a store hub associated with a virtual endless aisle, in response to an interaction of the customer with a product, emitting, by a smart shelf of the virtual endless aisle, an event, based on the event, transmitting, by the store hub to a display screen visible to the user, information concerning the product, displaying, by the display screen, a portion of the virtual endless aisle that includes the information concerning the product, detecting, by a camera associated with the virtual endless aisle, a product selection signal generated by the customer, and enabling, by the virtual endless aisle, the customer to purchase the product.

Abstract: A method for compressing and fusing sensor data. One example method includes collecting, by a sensor, data concerning an environment in which the sensor is located, compressing the data to generate compressed data, filtering the compressed data, fusing the compressed data with other data collected by another sensor to generate fused data, and transmitting and/or storing the fused data. The sensor may be a stereo camera, and the compressing may involve reducing a frame capture rate of the stereo camera.