Granted Patent B2

US 12,482,053 · App. 18/595,208 · Granted Nov 25, 2025

Inventors: Adnan M. Alattar (Tigard, OR); Ravi K. Sharma (Portland, OR); John A. Scriven (Oregon City, OR)

Assignee: Digimarc Corporation

G06T1/0028G06F18/22G06F21/16G06T1/005G06T1/0092G06V20/46G06V40/16G06V40/171G10L19/018

Abstract

Claims (93)

1 . A method for authenticating media content comprising:

reading a digital watermark embedded in media content;

obtaining an index to a block in a blockchain from a digital watermark payload obtained from the digital watermark;

requesting metadata from the blockchain based on the index;

extracting features from the media content to yield extracted features;

checking for matching between the extracted features and stored features in the metadata obtained from the block in the blockchain; and

determining alteration in the media content by:

detecting presence of valid digital watermarks in segments of the media content;

checking consistency between digital watermarks in different portions of the media content; and

checking a relationship of segment identifiers carried by digital watermarks in the segments.

2 . The method of claim 1 wherein said extracting features comprises:

computing robust hashes of segments of the media content; and

storing the robust hashes in the metadata in the blockchain.

3 . The method of claim 1 wherein extracting features comprises:

computing Mel Frequency Cepstrum (MFC) coefficients from audio segments of the media content; and

storing the MFC coefficients in the metadata in the blockchain.

4 . The method of claim 1 further comprising:

dividing the media content into a plurality of segments;

embedding a digital watermark carrying a unique segment identifier per segment in each of the plurality of segments; and

checking for presence of valid segment identifier in at least some of the plurality of segments to detect alteration.

5 . The method of claim 1 further comprising:

transcribing audio from the media content to generate a transcription;

storing the transcription in the metadata in the blockchain to yield a stored transcription; and

comparing a transcription of suspect audio against the stored transcription to detect alteration.

6 . The method of claim 1 further comprising:

computing facial feature locations from video frames of the media content;

storing the facial feature locations in the metadata in the blockchain, said storing yielding stored facial feature locations; and

comparing facial feature locations in suspect video frames against the stored facial feature locations to detect alteration.

7 . A method for authenticating media content comprising:

reading a digital watermark embedded in media content;

obtaining an index to a media content file from a digital watermark payload obtained from the digital watermark;

requesting the media content file based on the index;

checking for matching between extracted features and features in the media content file; and

determining whether the media content is an authorized derivative work by:

extracting an identifier of an original work from a blockchain associated with the media content;

retrieving blocks from a blockchain containing the identifier of the original work; and

verifying that a valid creation transaction exists in the blockchain linking the original work to the media content as an authorized derivative work.

8 . The method of claim 7 wherein the index comprises an index to a block in a blockchain, and the block in the blockchain provides an index to the media content file stored in a distributed storage system, wherein:

the distributed storage system divides the media content file into a plurality of chunks;

hashes the plurality of chunks to provide hashes that index the plurality of chunks in memory of the distributed storage system;

stores the plurality of chunks on different servers in the distributed storage system; and

provides a hash table storing the hashes that the distributed storage system uses as indexes to retrieve the plurality of chunks from the different servers.

9 . The method of claim 7 further comprising:

recording a creation transaction in the blockchain that links an original work to the media content file;

embedding an identifier of the original work in a digital watermark in the media content file; and

verifying authenticity by confirming presence of a creation transaction in the blockchain.

10 . The method of claim 7 further comprising:

assigning a unique instance identifier to a distributed instance of the media content file;

recording a distribution transaction in the blockchain using the unique instance identifier; and

tracking a distributed instance using a blockchain transaction.

11 . The method of claim 7 further comprising:

storing chunks of the media content file on different servers in a distributed storage system;

computing hashes of the chunks to form a hash tree;

storing a root hash of the hash tree in the blockchain; and

using the root hash to verify integrity of the media content file.

12 . The method of claim 7 wherein the index is associated with:

a block identifier from a blockchain transaction record; and

a distribution transaction.

13 . The method of claim 7 further comprising:

recording version control information in blockchain transactions to track derivative works created from the media content file.

14 . A system for authenticating media content comprising:

a watermark reader configured to read a digital watermark embedded in media content;

a blockchain interface configured to:

obtain an index to a block in a blockchain from a watermark payload obtained from the digital watermark;

request metadata from the blockchain based on the index;

a feature extractor configured to extract features from the media content; and

an authentication module configured to:

check for matching between the extracted features and stored features in the metadata obtained from the block in the blockchain;

detect presence of valid watermarks in segments of the media content;

check consistency between watermarks in different portions of the media content by verifying that interdependent payload portions between the different portions match; and

check a relationship of segment identifiers embedded as watermarks in the segments to detect alteration through frame removal, alteration or repetition.

15 . The system of claim 14 wherein the feature extractor is configured to:

compute robust hashes of segments of the media content; and

store the robust hashes in the metadata in the blockchain.

16 . The system of claim 14 wherein the authentication module is configured to:

detect presence of valid digital watermarks in segments of the media content;

check consistency between digital watermarks in different content tracks; and

check an order of segment identifiers embedded as digital watermarks.

17 . The system of claim 14 further comprising:

a distributed storage system configured to:

divide the media content into chunks;

store the chunks on different servers;

compute hashes of the chunks; and

provide a hash table for retrieving the chunks.

18 . The system of claim 14 further comprising:

a version control module configured to:

record creation and distribution of derivative works in the blockchain;

relate blocks of original and derivative works using identifiers; and

track relationships between original works and authorized derivatives.

19 . The system of claim 14 further comprising:

a transcription module configured to generate text transcriptions from audio content; wherein the authentication module is configured to compare transcriptions to detect audio alterations.

20 . The system of claim 14 wherein the feature extractor is configured to:

compute facial feature locations from video frames; wherein the authentication module is configured to compare facial feature locations to detect video alterations.

Assignments (3)

Continuity (3)

References Cited (89)