US 12,720,098 B1Granted

Methods and systems for identifying violation videos

OwnerPierre Raczinventor

Life of this patent

Estimating the term…

Loading patent…

Abstract

A method of identifying a user associated with a violation video may include the following steps: retrieving a source video; applying one or more watermarks to the source video, wherein the one or more watermarks encode viewing information related to the source video; displaying the source video and the one or more watermarks to the user; retrieving the violation video; identifying a region of interest in the violation video; identifying a section of the source video which corresponds to the violation video; registering the region of interest with a corresponding region in the source video; determining a transfer function relating pixels in the region of interest of the violation video to pixels in the corresponding region of the source video; determining expected values of the pixels in the region of interest of the violation video; comparing the expected values of the pixels in the region of interest to actual values of the pixels in the region of interest; identifying features of one or more of the one or more watermarks; and recovering the viewing information from the identified features of the one or more watermarks.

Claims18

Dependent claims

A method of identifying a source of leaked security footage encoded in viewing information associated with a violation video captured at a secure and uniquely identified security system monitoring workstation, the method comprising:

collecting and storing video streams from a network of security cameras, each observing a location for security purposes, at a security system video archive server,
authenticating, at said security system video archive server, a user or a user's security system monitoring workstation to receive a source video from said security system video archive server, said source video being a security recording of said location and being captured by at least one of said security cameras, said viewing information having a correlation with said user or said user's security system monitoring workstation, said security system monitoring workstation allowing an authenticated user to view security videos from any one of said network of security cameras;
applying one or more watermarks to the source video, wherein the one or more watermarks encode viewing information related to the source video and to said authenticated user or user's security system monitoring workstation, said one or more watermarks comprising watermarked portions present on spatial regions distributed spatially across the source video;
displaying a displayed video comprising the source video and the one or more watermarks to the user;
retrieving the violation video, said violation video being a recording of the displayed video at the security system monitoring workstation;
identifying features of one or more of the one or more watermarks in the violation video;
and
recovering the viewing information from the identified features of the one or more watermarks;
wherein said identifying said features comprises:
identifying a region of interest in the violation video;
identifying a section of the source video which corresponds to the violation video;
registering the region of interest with a corresponding region in the source video;
determining a transfer function relating pixels in the region of interest of the violation video to pixels in the corresponding region of the source video using multiple frames of the violation video, the transfer function accounting for glare and/or shadows in the violation video not present in the source video;
determining, using the transfer function, expected values of the pixels in the region of interest of the violation video, said expected values representing an approximation of said displayed video by removing the identified glare and/or shadows introduced in said violation video;
comparing the expected values of the pixels in the region of interest to actual values of the pixels in the region of interest of the source video.

12 more elements

A method of extracting information from a violation video, the method comprising:

obtaining a violation video and a source video which corresponds to the violation video, said violation video being acquired by a video camera capturing a display screen displaying a displayed video, said displayed video comprising said source video combined with one or more watermarks, said violation video comprising irregularities including glare and/or shadows resulting from said capturing that are not said one or more watermarks;
identifying a region of interest in the violation video;
registering the region of interest with a corresponding region in the source video;
identifying glare and/or shadows in the region of interest of the violation video not present in the source video using multiple frames of the violation video;
determining, using the identified glare and/or shadows, expected values of the pixels in the region of interest of the violation video, said expected values representing an approximation of said displayed video by removing the identified glare and/or shadows introduced in said violation video;
comparing the expected values of the pixels in the region of interest to actual values of the pixels in the region of interest from said source video to identify features of one or more watermarks based on differences between the expected values of the pixels in the region of interest and the actual values of the pixels in the region of interest from said source video.

3 more elements

Dependencies are inferred from the claim text.

Facts

Application: 17/083,772Filed 29 Oct 2020 · us provisional application
Priority: 29 Oct 2019
Granted: 25 Aug 2026
Term ends: Not estimatedThe source records don’t support an estimate.

Classification10 codes

H04N 19/467
Embedding additional information in the video signal during the compression process: characterised by the embedded information being invisible,…
G06T 1/0085
Image watermarking: time domain based watermarking, e.g. watermarks spread over several images
G06V 20/52
Context or environment of the image: surveillance or monitoring of activities, e.g. for recognising suspicious objects
G06F 21/31
Authentication, i.e. establishing the identity or authorisation of security principals: user authentication