Abstract: A goal of the disclosure is to provide real-time adjustment of a deep learning-based tracking system to track a moving individual without using a labeled set of training data. Disclosed are systems and methods for tracking a moving individual with an autonomous drone. Initialization video data of the specific individual is obtained. Based on the initialization video data, real-time training of an input neural network is performed to generate a detection neural network that uniquely corresponds to the specific individual. Real-time video monitoring data of the specific individual and the surrounding environment is captured. Using the detection neural network, target detection is performed on the real-time video monitoring data and a detection output corresponding to a location of the specific individual within a given frame of the real-time video monitoring data is generated.

Abstract: Disclosed are technologies for autonomous tracking. An initial coordinate of a beacon device carried by a user is registered as a dead reckoning waypoint with a drone configured to track the user. The drone receives IMU measurements from the beacon as the user moves. For each IMU measurement, a displacement vector characterizing user movement is calculated. Estimated beacon locations are calculated by dead reckoning, based on the displacement vectors and the dead reckoning waypoint. Later, an updated dead reckoning waypoint is calculated by obtaining the current location coordinate of the drone and performing optical triangulation to determine a relative position of the user with respect to the drone. The updated dead reckoning waypoint does not depend on previously estimated beacon locations, and accumulated IMU/estimation error is eliminated. Tracking continues, where subsequent estimated locations of the beacon are calculated by dead reckoning based on the updated dead reckoning waypoint.

Abstract: First and second screenshot images are obtained from a monitoring application provided on a first computing device. Each screenshot image comprises a plurality of content portions displayed by a communication application on the first computing device, and content boxing is performed to calculate a plurality of content boxes for the plurality of content portions. Each content box is classified as containing textual communication content or image communication content. Textual communications are extracted via Optical Character Recognition (OCR) and object identifiers are extracted from the image communications via image recognition. At least one shared content box present in both the first and second screenshot images is identified and used to temporally align the extracted textual communications. The temporally aligned textual communications are condensed into a textual communication sequence.

Abstract: A drone receives an activation command indicating a user's need for monitoring, and is deployed based on the activation command and a set of initial operational parameters. The drone autonomously navigates to a first position with respect to the user and performs a first configured action. A plurality of monitoring data signals corresponding to the user and surrounding environment is captured using sensors on the drone, and is wirelessly transmitted by the drone to a remote monitoring system. The monitoring data signals are continuously analyzed to generate updated operational parameters causing the drone to autonomously navigate to a second position and perform a second configured action. A third configured action is received by the drone from the remote monitoring system, wherein the third configured action is generated based on a threat analysis performed by the remote monitoring system on the monitoring data signals.