US 11,676,386 B2
Method and system for automated analysis of human behavior utilizing video data
Aadalarasan Bhavani Saravanan, Mumbai (IN); Christen Miller, Mumbai (IN); Ryan Dsouza, Mumbai (IN); Chitrangada Patra, Mumbai (IN); Dinesh Avula, Mumbai (IN); Ankit Ratan, Mumbai (IN); Arpit Ratan, Mumbai (IN); and Ankur Pandey, Mumbai (IN)
Assigned to Signzy Technologies Private Limited, Mumbai (IN)
Filed by Signzy Technologies Private Limited, Mumbai (IN)
Filed on Dec. 17, 2020, as Appl. No. 17/125,983.
Prior Publication US 2022/0027635 A1, Jan. 27, 2022
Int. Cl. G06V 20/00 (2022.01); G06V 20/52 (2022.01); G06N 20/00 (2019.01); H04N 7/18 (2006.01); G06V 40/20 (2022.01)
CPC G06V 20/52 (2022.01) [G06N 20/00 (2019.01); G06V 40/20 (2022.01); H04N 7/181 (2013.01)] 14 Claims
OG exemplary drawing
 
1. A computer-implemented method for automated analysis of human behavior, wherein the automated analysis of human behavior is performed for determining fraudulent behavior of one or more users, the computer-implemented method comprising:
collecting, with a fraudulent behavior detection system that includes a processor, a technical data and a video data from one or more data sources and one or more video sources, wherein the one or more data sources and the one or more video sources are associated with the fraudulent behavior detection system;
training, with the fraudulent behavior detection system that includes the processor, the fraudulent behavior detection system with the collected technical data and the video data in real-time, wherein the fraudulent behavior detection system facilitates training in real-time, wherein the training of the fraudulent behavior detection system is performed using one or more hardware-run machine learning algorithms and one or more hardware-run deep learning algorithms;
receiving, with the fraudulent behavior detection system that includes the processor, a live video stream data from the one or more video sources, wherein the live video stream data is received continuously in real-time, wherein the one or more video sources are installed at a facility;
analyzing, with the fraudulent behavior detection system that includes the processor, the live video stream data received from the one or more video sources installed at the facility in real-time, wherein the live video stream data is analyzed using the one or more hardware-run machine learning algorithms and the one or more hardware-run deep learning algorithms;
predicting, with the fraudulent behavior detection system that includes the processor, likelihood of the fraudulent behavior of humans based on the analysis of the live video stream data, wherein the prediction is performed for alarming concerned authorities of the facility about likelihood of fraudulent behavior;
scanning, with the fraudulent behavior detection system that includes the processor, the live video stream data, wherein the scanning of the live video stream data is performed for dividing the live video stream data into one or more frames of images, wherein the frames of images are analyzed for predicting the likelihood of the fraudulent behavior; and
mapping, with the fraudulent behavior detection system that includes the processor, moving pattern of humans from the live video stream data to moving pattern of fraudsters in the video data stored in a database, wherein the mapping is performed to identify fraudsters from humans in the live video stream data.