Method and apparatus for authenticating video content

Patent metadata

Field	Value
Publication number	US-8989376-B2
Application number	US-201213434399-A
Country	US
Kind code	B2
Filing date	Mar 29, 2012
Priority date	Mar 29, 2012
Publication date	Mar 24, 2015
Grant date	Mar 24, 2015

Abstract

A method for authenticating video content includes: receiving a digital signature, an unsecured video fingerprint, and an unsecured video content from a transmitting node at a receiving node in a communication network; determining if the digital signature is consistent with the unsecured video fingerprint at the receiving node to verify the unsecured video fingerprint; and determining if the unsecured video fingerprint is consistent with the unsecured video content at the receiving node to verify the unsecured video content in a manner that tolerates a predetermined measure of loss in the unsecured video content. If the unsecured video fingerprint and the unsecured video content are verified, the unsecured video content is authenticated for subsequent use at the receiving node. A receiving node associated with the method includes an input module, a fingerprint verification module, a content verification module, and a controller module.

First claim

We claim:

1. A method for authenticating video content, comprising:
receiving a digital signature, a received video fingerprint, and a video content from a transmitting node at a receiving node in a communication network, wherein the received video fingerprint is a received version of an original video fingerprint that is derived from an original video content using a fingerprinting algorithm that derives said original video fingerprint based on trajectories of movement of salient feature points detected from sampled frames of the original video content prior to transmission of the original video fingerprint by the transmitting node;
verifying the received video fingerprint by determining if the digital signature is consistent with the received video fingerprint at the receiving node; and
verifying the video content by determining if the received video fingerprint is consistent with the video content at the receiving node in a manner that tolerates a predetermined measure of loss in the video content, wherein the verifying the video content includes generating a fresh video fingerprint by processing the video content at the receiving node using the fingerprinting algorithm;
wherein, if the received video fingerprint and the video content are verified, the video content is authenticated for subsequent use at the receiving node.

2. The method of claim 1 wherein the manner that tolerates a predetermined measure of loss in the video content is based on a distance metric between the received video fingerprint and the fresh video fingerprint at the receiving node using a complexity-invariant distance measure algorithm.

3. The method of claim 1 wherein the digital signature is produced from an original hash value using an encryption algorithm and a private key prior to transmission of the digital signature by the transmitting node;
wherein the original hash value is derived from the original video fingerprint using a hashing algorithm prior to encryption of the original hash value.

4. The method of claim 3 , in conjunction with verifying the received video fingerprint, the method further comprising:
decrypting the digital signature using a decryption algorithm and a public key at the receiving node to obtain a decrypted hash value relating to the original hash value;
processing the received video fingerprint using the hashing algorithm at the receiving node to obtain a fresh hash value relating to the original hash value; and
comparing the fresh hash value to the decrypted hash value at the receiving node such that the received video fingerprint is verified if the fresh hash value matches the decrypted hash value.

5. The method of claim 1 , in conjunction with verifying the video content, the method further comprising:
generating the fresh video fingerprint by processing the video content at the receiving node using the fingerprinting algorithm;
determining a distance metric between the received video fingerprint and the fresh video fingerprint at the receiving node using a complexity-invariant distance measure algorithm; and
comparing the distance metric to a predetermined threshold at the receiving node such that the video content is verified if the distance metric does not exceed the predetermined threshold.

6. The method of claim 5 , in conjunction with using the fingerprinting algorithm, the method further comprising:
selecting the sample of video frames from the video content and arranging the sample video frames in a concatenated time sequence;
detecting the salient feature points in each sample video frame;
computing angular orientations of optical flow for each salient feature point in each sample video frame in relation to the corresponding salient feature point in the next sample video frame of the concatenated time sequence;
distributing the angular orientations for the salient feature points of each sample video frame into corresponding angular range bins for each sample video frame;
concatenating the values in each angular range bin for the sample video frames over the concatenated time sequence to form a histogram for each angular range bin; and
normalizing the set of histograms for the angular range bins to form a corresponding set of motion time series that establish the fresh video fingerprint.

7. The method of claim 6 , in conjunction with using the fingerprinting algorithm to establish the fresh video fingerprint, the method further comprising:
compressing each motion time series using a linear segmentation algorithm to convert the corresponding histogram into a corresponding sequence of linear segments; and
extracting major inclines from each compressed motion time series based at least in part on selecting linear segments that are greater than a predetermined threshold value for at least one of a time characteristic and an amplitude characteristic to form a corresponding set of motion time series for the fresh video fingerprint represented by the extracted major inclines.

8. The method of claim 5 wherein the original video fingerprint, the received video fingerprint, and fresh video fingerprint each comprise a corresponding set of motion time series formed by reducing corresponding histograms to sequences of linear segments and extracting major inclines from the sequences of linear segments.

9. The method of claim 8 , in conjunction with using the complexity-invariant distance measure algorithm, the method further comprising:
pairing each motion time series of the received video fingerprint with a corresponding motion time series of the fresh video fingerprint;
aligning each paired motion time series based at least in part on identification of similar major inclines in the corresponding paired motion time series; and
determining a distance measure between each aligned motion time series using the complexity-invariant distance measure algorithm.

10. The method of claim 5 wherein the original video fingerprint, the received video fingerprint, and the fresh video fingerprint each comprise a corresponding set of motion time series formed by corresponding histograms.

11. The method of claim 10 , in conjunction with using the complexity-invariant distance measure algorithm, the method further comprising:
compressing each motion time series of the received video fingerprint using a linear segmentation algorithm to convert the corresponding histogram into a corresponding sequence of linear segments;
extracting major inclines from each compressed motion time series of the received video fingerprint based at least in part on selecting linear segments that are greater than a predetermined threshold value for at least one of a time characteristic and an amplitude characteristic to form a corresponding set of motion time series for the received video fingerprint represented by the extracted major inclines;
compressing each motion time series of the fresh video fingerprint using the linear segmentation algorithm to convert the corresponding histogram into a corresponding sequence of linear segments;
extracting major inclines from each compressed motion time series of the fresh video fingerprint based at least in part on selecting linear segments that are greater than the predetermined threshold value for at least one of the time characteristic and the amplitude characteristic to form a corresponding set of motion time series for the fresh video fingerprint represented by the extracted major inclines;
pairing each motion time series of the received video fingerprint with a corresponding motion time series of the fresh video fingerprint;
aligning each paired motion time series based at least in part on identification of similar major inclines in the corresponding p

Assignees

• Ren Yansong
• O'Gorman Lawrence
• Zhang John R
• Wood Thomas L
• Alcatel Lucent

Inventors

• Ren Yansong
• O'Gorman Lawrence
• Zhang John R
• Wood Thomas L

Classifications

• H04L9/3247 Primary

  involving digital signatures · CPC title

• H04L63/12 Primary

  Applying verification of the received information (cryptographic mechanisms or cryptographic arrangements for data integrity or data verification H04L9/32) · CPC title

• H04L63/123

  received data contents, e.g. message integrity · CPC title

• H04L9/00

  {Cryptographic mechanisms or cryptographic} arrangements for secret or secure communications; Network security protocols · CPC title

• H04L2209/608

  Watermarking · CPC title