# US 2022/0020383 A1 · Method for adding watermark information, method for extracting watermark information, and electronic device

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2022/0020383 A1
- Document ID: US20220020383A1
- Status: Published application
- Kind: A1
- Application number: 17/489,603
- Filed: 2021-09-29
- Published: 2022-01-20
- Earliest claimed: 2020-02-04
- Assignees: Beijing Dajia Internet Information Technology Co., Ltd.
- Applicants: Beijing Dajia Internet Information Technology Co., Ltd.
- Inventors: Chen ZHANG; Xiguang ZHENG; Liang GUO
- Web page: https://patentgrep.com/patent/US20220020383A1

## CPC

- G10L 19/018: Audio watermarking, i.e. embedding inaudible data in the audio signal
- G10L: SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING

## Abstract

Provided is a method for adding watermark information. The method includes: acquiring M first audio signal frames in a first audio signal; acquiring N watermark information items in watermark information; determining M*N adding parameters; acquiring M second audio signal frames added with the watermark information based on the M*N adding parameters; and determining a second audio signal based on the M second signal frames added with the watermark information.

## Claims

### Claim 1

A method for adding watermark information, executed by an electronic device, the method comprising:
acquiring M first audio signal frames in a first audio signal, where M is a positive integer larger than 1;
acquiring N watermark information items in watermark information, where N is a positive integer larger than 1;
determining M*N adding parameters, wherein each of the adding parameters corresponds to one of the watermark information items and one of the first audio signal frames;
acquiring M second audio signal frames added with the watermark information based on the M*N adding parameters, wherein the second audio signal frame added with the watermark information is acquired by adding the N watermark information items to the first audio signal frame based on N adding parameters, wherein the N adding parameters correspond to the first audio signal frame and correspond to N watermark information items; and
determining a second audio signal based on the M second signal frames added with the watermark information.

### Claim 2 (depends on 1)

The method according to claim 1, wherein the adding parameter comprises a target position and an information strength; and said acquiring M second audio signal frames added with the watermark information based on the M*N adding parameters comprises:
acquiring the second audio signal frame added with the watermark information by adding each of the N watermark information items in the first audio signal frame based on the target position and the information strength.

### Claim 3 (depends on 1)

The method according to claim 1, wherein said acquiring M second audio signal frames added with the watermark information based on the M*N adding parameters comprises:
acquiring parameter information of the first audio signal frame, wherein the parameter information comprises at least one of amplitude information or phase information; and
acquiring the second audio signal frame added with the watermark information by adjusting the parameter information of the first audio signal frame based on the N watermark information items and the N adding parameters corresponding to the first audio signal frame.

### Claim 4 (depends on 1)

The method according to claim 1, further comprising:
acquiring the first audio signal by transforming a third audio signal;
wherein the third audio signal is a time domain audio signal, and the first audio signal is a time-frequency domain audio signal.

### Claim 5 (depends on 1)

The method according to claim 1, wherein said acquiring the N watermark information items in the watermark information comprises:
acquiring converted watermark information by performing binary conversion on the watermark information; and
determining the N watermark information items based on N bits in the converted watermark information, wherein each bit corresponds to one watermark information item.

### Claim 6 (depends on 5)

The method according to claim 5, wherein said acquiring the converted watermark information by performing binary conversion on the watermark information comprises:
acquiring binary watermark information by performing the binary conversion on the watermark information; and
determining converted watermark information corresponding to the binary watermark information according to a reference conversion relationship, wherein the reference conversion relationship comprises converted binary numbers corresponding to original binary numbers.

### Claim 7 (depends on 1)

The method according to claim 1, wherein the adding parameter comprises a target position;
said acquiring M second audio signal frames added with the watermark information based on the M*N adding parameters comprises:
adjusting the first audio signal frame by using the following formula:
{
P
w
⁡
(
n
,
k
)
=
P
⁡
(
n
,
k
)
·
Mask
b
⁡
(
n
,
k
)
·
x
,
if
⁢
⁢
I
⁡
(
b
)
=
1
P
w
⁡
(
n
,
k
)
=
P
⁡
(
n
,
k
)
·
Mask
b
⁡
(
n
,
k
)
/
y
,
if
⁢
⁢
I
⁡
(
b
)
=
0
;
wherein n represents the first audio signal frame, k represents a central frequency of the first audio signal frame, P (n,k) represents parameter information of the first audio signal frame, Pw(n,k) represents the parameter information of the second audio signal frame added with the watermark information, I (b) represents a bth watermark information item in the watermark information, Maskb(n,k) represents the target position corresponding to the bth watermark information item in the audio signal frame, b represents a positive integer, and x and y represent reference values.

### Claim 8 (depends on 2)

The method according to claim 2, wherein said acquiring the second audio signal frame added with the watermark information by adding each of the N watermark information items in the first audio signal frame based on the target position and the information strength comprises:
adjusting the first audio signal frame by using the following formula:
{
P
w
⁡
(
n
,
k
)
=
P
⁡
(
n
,
k
)
·
Mask
b
⁡
(
n
,
k
)
·
10
s
b
20
,
if
⁢
⁢
I
⁡
(
b
)
=
1
P
w
⁡
(
n
,
k
)
=
P
⁡
(
n
,
k
)
·
Mask
b
⁡
(
n
,
k
)
/
10
s
b
20
,
if
⁢
⁢
I
⁡
(
b
)
=
0
;
wherein n represents the first audio signal frame, k represents a central frequency of the first audio signal frame, P(n,k) represents parameter information of the first audio signal frame, Pw (n,k) represents the parameter information of the second audio signal frame added with the watermark information, I (b) represents a bth watermark information item in the watermark information, Maskb (n,k) represents the target position corresponding to the bth watermark information item in the audio signal frame, and sb represents the information strength corresponding to the bth watermark information item in the audio signal frame.

### Claim 9 (depends on 1)

The method according to claim 1, wherein said determining M*N adding parameters comprises:
encrypting the watermark information according to a reference key corresponding to the watermark information; and
determining the M*N adding parameters based on the encrypted watermark information and a reference function.

### Claim 10

A method for extracting watermark information, executed by an electronic device, the method comprising:
acquiring a second audio signal added with watermark information;
determining N adding parameters in a second audio signal frame of the second audio signal, wherein each of the adding parameters corresponds to one watermark information item in the watermark information, and N is a positive integer;
acquiring N decoded watermark information items, wherein one decoded watermark information item corresponds to one watermark information item; and
extracting watermark information from the second audio signal frame based on the N adding parameters and the N decoded watermark information items.

### Claim 11 (depends on 10)

The method according to claim 10, wherein the adding parameter comprises a target position and an information strength; and said extracting watermark information from the second audio signal frame based on the N adding parameters and the N decoded watermark information items comprises:
extracting the watermark information from the second audio signal frame based on the target positions and information strengths of the N adding parameters in the second audio signal frame and the N decoded watermark information items.

### Claim 12 (depends on 10)

The method according to claim 10, wherein the adding parameter comprises a target position; and said extracting watermark information from the second audio signal frame based on the N adding parameters and the N decoded watermark information items comprises:
acquiring parameter information of the second audio signal frame, wherein the parameter information comprises at least one of amplitude information or phase information;
acquiring a plurality of target parameter information in the second audio signal frame based on the target positions of the N adding parameters; and
extracting the watermark information from the plurality of target parameter information based on the N adding parameters and the N decoded watermark information items.

### Claim 13 (depends on 12)

The method according to claim 12, wherein said acquiring the plurality of target parameter information in the second audio signal frame based on the target positions of the N adding parameters comprises:
acquiring a plurality of converted parameter information in the second audio signal frame based on the target positions of the N adding parameters; and
determining the plurality of target parameter information according to a reference conversion relationship and the plurality of converted parameter information, wherein each piece of the target parameter information and the converted parameter information is binary information, and the reference conversion relationship comprises converted binary numbers corresponding to original binary numbers.

### Claim 14 (depends on 10)

The method according to claim 10, wherein the adding parameter comprises a target position; and said extracting watermark information from the second audio signal frame based on the N adding parameters and the N decoded watermark information items comprises:
acquiring a plurality of target parameter information in the second audio signal frame based on the target position of each of the watermark information items in the second audio signal frame;
determining relevancy of watermark information items corresponding to any two pieces of target parameter information adjacent to each other based on the any two pieces of target parameter information and two of the decoded watermark information items corresponding to the any two pieces of target parameter information; and
extracting the watermark information items corresponding to the any two pieces of target parameter information from the second audio signal frame based on the relevancy.

### Claim 15 (depends on 14)

The method according to claim 14, wherein said determining the relevancy of the watermark information items corresponding to the any two pieces of target parameter information adjacent to each other based on the any two pieces of target parameter information and the two of the decoded watermark information items corresponding to the any two pieces of target parameter information comprises:
determining the relevancy by using the following formula:
C=Pwe,f·We,f;
wherein C represents the relevancy, Pwe,f represents target parameter information acquired by combining target parameter information corresponding to an eth watermark information item and target parameter information corresponding to an fth watermark information item, We,f represents a decoded watermark information item acquired by combining two of the decoded watermark information items corresponding to Pwe,f, and the eth watermark information item and the fth watermark information item represent any two watermark information items adjacent to each other.

### Claim 16 (depends on 14)

The method according to claim 14, wherein said extracting the watermark information items corresponding to the any two pieces of target parameter information from the second audio signal frame based on the relevancy comprises:
extracting watermark information items 1 from the second audio signal frame in response to the relevancy being a first reference value; or
extracting watermark information items 0 from the second audio signal frame in response to the relevancy being a second reference value.

### Claim 17 (depends on 14)

The method according to claim 14, wherein the adding parameter further comprises an information strength; and said extracting watermark information from the second audio signal frame based on the N adding parameters and the N decoded watermark information items comprises:
determining the relevancy corresponding to the watermark information items by using the following formula:
C=Pwe,f·We,f=(Pe,f±We,f)·We,f=Pe,f·We,f+(n+m)s2;
wherein n represents a quantity of target positions corresponding to an eth watermark information item, m represents a quantity of target positions corresponding to an fth watermark information item, s represents an information strength of the eth watermark information item and the fth watermark information item, Pe,f represents parameter information acquired by combining parameter information corresponding to the eth watermark information item and parameter information corresponding to the ft watermark information item before the watermark information is added;
extracting watermark information items 1 from the second audio signal frame in response to

C
(
n
+
m
)
⁢
s
2

being not less than a reference threshold and the relevancy being a first reference value; or
extracting watermark information items 0 from the second audio signal frame in response to

C
(
n
+
m
)
⁢
s
2

being not less than the reference threshold and the relevancy being a second reference value.

### Claim 18 (depends on 17)

The method according to claim 17, further comprising:
extracting watermark information items from the second audio signal frame based on the relevancy and confidence in response to

C
(
n
+
m
)
⁢
s
2

being less than the reference threshold, wherein the confidence represents credibility of the watermark information items extracted based on the relevancy.

### Claim 19 (depends on 10)

The method according to claim 10, wherein said determining N adding parameters in the second audio signal frame comprises:
acquiring decrypted watermark information by decrypting the watermark information according to a reference key corresponding to the watermark information; and
determining the N adding parameters according to the reference key and a reference function.

### Claim 20

An electronic device, comprising:
at least one processor; and
a volatile or nonvolatile memory configured to store at least one instruction executable by the at least one processor;
wherein the at least one processor, when executing the at least one instruction, is caused to perform:
acquiring M first audio signal frames in a first audio signal, where M is a positive integer larger than 1;
acquiring N watermark information items in watermark information, where N is a positive integer larger than 1;
determining M*N adding parameters, wherein each of the adding parameters corresponds to one of the watermark information items and one of the first audio signal frames;
acquiring M second audio signal frames added with the watermark information based on the M*N adding parameters, wherein the second audio signal frame added with the watermark information is acquired by adding the N watermark information items to the first audio signal frame based on N adding parameters, wherein the N adding parameters correspond to the first audio signal frame and correspond to N watermark information items; and
determining a second audio signal based on the M second signal frames added with the watermark information
