Search

CN-122023098-A - Watermark information detection method, device, program product and electronic equipment

CN122023098ACN 122023098 ACN122023098 ACN 122023098ACN-122023098-A

Abstract

The application discloses a watermark information detection method, a device, a program product and electronic equipment, and relates to the field of financial science and technology or other related technical fields, wherein the method comprises the steps of detecting the similarity between each preset image in L preset images and an image to be detected; and under the condition that the maximum value of the L similarities corresponding to the L preset images is greater than or equal to the preset similarity, taking the preset image corresponding to the maximum value as a target image, performing watermark extraction operation on the target image to obtain preset watermark information corresponding to the target image, and taking the preset watermark information corresponding to the target image as target watermark information of the image to be detected. The application solves the technical problem of low efficiency of watermark detection on the rotated and cut image based on the prior art.

Inventors

  • WANG LIBIAO

Assignees

  • 中国工商银行股份有限公司

Dates

Publication Date
20260512
Application Date
20260119

Claims (10)

  1. 1. A method for detecting watermark information, comprising: Detecting the similarity between each preset image in L preset images and an image to be detected, wherein L is a positive integer, the image to be detected is an image subjected to clipping/rotation processing, and each preset image is an image with embedded preset watermark information stored in a preset database of a financial institution; taking a preset image corresponding to the maximum value as a target image under the condition that the maximum value in the L similarities corresponding to the L preset images is larger than or equal to the preset similarity; Watermark extraction operation is carried out on the target image, and preset watermark information corresponding to the target image is obtained; And taking the preset watermark information corresponding to the target image as target watermark information of the image to be detected, wherein the target watermark information is used for representing the personnel identification of the sender of the image to be detected.
  2. 2. The method for detecting watermark information according to claim 1, wherein before detecting the similarity between each of the L preset images and the image to be detected, the method for detecting watermark information further comprises: Obtaining L original images to be embedded with the preset watermark information, wherein the security level corresponding to each original image is greater than or equal to a preset security level; Performing coding operation on the preset watermark information to obtain an initial coding sequence corresponding to the preset watermark information; performing coding enhancement on the initial coding sequence to obtain a target coding sequence corresponding to the preset watermark information; And generating the L preset images based on the target coding sequence and the L original images.
  3. 3. The method according to claim 2, wherein generating the L preset images based on the target coding sequence and the L original images includes: storing transparency information of each original image in the L original images into a preset file, wherein the transparency information at least comprises a transparency value of each pixel in each original image; Performing color space transformation operation on each original image to obtain a first image corresponding to each original image, wherein the color information of each original image is represented by a three-primary-color channel, and the color information in each first image is represented by a chromaticity channel and a brightness channel; performing image capturing operation on each first image based on a preset proportion to obtain second images corresponding to each first image, wherein each second image is a local image positioned at the central position of each first image; Performing block division on each second image to obtain M blocks corresponding to each second image, wherein M is a positive integer; embedding the target coding sequence into a brightness channel corresponding to each block based on preset embedding strength corresponding to each block in the M blocks to obtain M target blocks corresponding to the M blocks; And combining M target blocks corresponding to each second image in the L second images to obtain L preset images corresponding to the L second images.
  4. 4. The watermark information detection method according to claim 3, wherein before embedding the target coding sequence into the luminance channel corresponding to each block based on the preset embedding strength corresponding to each block of the M blocks, the watermark information detection method further comprises: Collecting global variances corresponding to each second image and local variances corresponding to each block in each second image, wherein each global variance is used for representing brightness change information of pixels in each second image, and each local variance is used for representing brightness change information of pixels in each block; and determining the preset embedding strength corresponding to each block based on the global variance corresponding to each second image and the local variance corresponding to each block in each second image.
  5. 5. The method for detecting watermark information according to claim 3, wherein embedding the target coding sequence into the luminance channel corresponding to each block based on a preset embedding strength corresponding to each block of the M blocks, to obtain M target blocks corresponding to the M blocks, comprises: Performing linear transformation operation on the brightness value of each block on a brightness channel to obtain a frequency domain coefficient matrix corresponding to each block, wherein the linear transformation operation is used for transforming the brightness value of each pixel on the brightness channel to a frequency domain; Based on the preset embedding strength corresponding to each block in the M blocks, embedding the target coding sequence into a preset low-frequency coefficient in a frequency domain coefficient matrix corresponding to each block to obtain a target coefficient matrix corresponding to each block; Performing inverse transformation operation on the target coefficient matrix corresponding to each block to obtain a target brightness value of each block on a brightness channel, wherein the bit value embedded in the target coding sequence in the target brightness value is inverse operation of the linear transformation operation; And combining the target brightness value of each block on the brightness channel, the chromaticity value on the chromaticity channel and the transparency value in the preset file to obtain M target blocks corresponding to the M blocks.
  6. 6. The method for detecting watermark information according to claim 1, wherein performing watermark extraction operation on the target image to obtain preset watermark information corresponding to the target image, comprises: performing block division on the target image to obtain M target blocks corresponding to the target image; Performing linear transformation operation on the brightness value of each target block on a brightness channel to obtain a target coefficient matrix corresponding to each target block; Counting symbol information corresponding to a target coefficient matrix corresponding to each target block, wherein the symbol information is used for representing coefficient symbols of preset low-frequency coefficients in the target coefficient matrix; Determining a target coding sequence corresponding to the target image based on symbol information corresponding to a target coefficient matrix corresponding to each target block; and decoding the target coding sequence corresponding to the target image to obtain the preset watermark information corresponding to the target image.
  7. 7. The method for detecting watermark information according to claim 2, wherein after performing an encoding operation on the preset watermark information to obtain an initial encoding sequence corresponding to the preset watermark information, the method for detecting watermark information further comprises: encrypting the initial coding sequence to obtain an encrypted coding sequence corresponding to the preset watermark information; Adding a preset prefix mark to the encryption coding sequence to obtain a file coding sequence corresponding to the preset watermark information; And storing the file coding sequence to a preset position in metadata corresponding to a preset data file to obtain target watermark metadata corresponding to the preset data file, wherein the preset data file is a text/audio/video file in which preset watermark information needs to be embedded.
  8. 8. A watermark information detection apparatus, comprising: The device comprises a similarity detection unit, a watermark detection unit and a watermark detection unit, wherein the similarity detection unit is used for detecting the similarity between each preset image and an image to be detected in L preset images, L is a positive integer, the image to be detected is an image subjected to clipping/rotation processing, and each preset image is an image embedded with preset watermark information and stored in a preset database of a financial institution; a target image determining unit, configured to, when a maximum value of L similarities corresponding to the L preset images is greater than or equal to a preset similarity, take the preset image corresponding to the maximum value as a target image; The watermark extraction unit is used for carrying out watermark extraction operation on the target image to obtain preset watermark information corresponding to the target image; The information determining unit is used for taking preset watermark information corresponding to the target image as target watermark information of the image to be detected, wherein the target watermark information is used for representing personnel identification of a sender of the image to be detected.
  9. 9. A computer program product, characterized in that the computer program product comprises a computer program, wherein the computer program product is controlled to perform the method of detecting watermark information according to any of claims 1 to 7 when the computer program is run.
  10. 10. An electronic device comprising one or more processors and a memory for storing one or more programs, wherein the one or more programs, when executed by the one or more processors, cause the one or more processors to implement the method of detecting watermark information of any of claims 1 to 7.

Description

Watermark information detection method, device, program product and electronic equipment Technical Field The present application relates to the field of financial science and technology or other related technical fields, and in particular, to a method, an apparatus, a program product, and an electronic device for detecting watermark information. Background With the rapid development of information technology, unstructured data (such as images, audio, video and documents) play an increasingly important role in the financial field as a carrier of information transmission, and accordingly, how to implement information transmission tracing and security protection of unstructured data is a problem to be solved in the current financial field, in the prior art, although various blind watermarking technologies are used for security protection of image data, the technologies generally face the following challenges: Although the traditional blind watermarking technology can resist image compression, noise pollution and attacks in certain spatial domains to a certain extent, when images subjected to geometric transformation operations such as rotation, clipping and the like are detected, the detection efficiency of the watermark is reduced, and even watermark information cannot be successfully recovered sometimes, so that the usability and the protection effect of the watermarking technology are weakened. In view of the above problems, no effective solution has been proposed at present. Disclosure of Invention The application provides a watermark information detection method, device, program product and electronic equipment, which at least solve the technical problem of low watermark detection efficiency of rotated and cut images based on the prior art. According to one aspect of the application, a watermark information detection method is provided, and the watermark information detection method comprises the steps of detecting similarity between each preset image in L preset images and an image to be detected, wherein L is a positive integer, the image to be detected is an image subjected to clipping/rotation processing, each preset image is an image embedded with preset watermark information stored in a preset database of a financial institution, when the maximum value of the L similarities corresponding to the L preset images is greater than or equal to the preset similarity, the preset image corresponding to the maximum value is taken as a target image, watermark extraction operation is conducted on the target image to obtain preset watermark information corresponding to the target image, the preset watermark information corresponding to the target image is taken as target watermark information of the image to be detected, and the target watermark information is used for representing personnel identification of a sender of the image to be detected. Optionally, before detecting the similarity between each preset image in the L preset images and the image to be detected, the watermark information detection method further comprises the steps of obtaining L original images to be embedded with preset watermark information, wherein the security level corresponding to each original image is greater than or equal to the preset security level, performing coding operation on the preset watermark information to obtain an initial coding sequence corresponding to the preset watermark information, performing coding enhancement on the initial coding sequence to obtain a target coding sequence corresponding to the preset watermark information, and generating L preset images based on the target coding sequence and the L original images. The method comprises the steps of generating L preset images based on a target coding sequence and L original images, storing transparency information of each original image in the L original images to a preset file, wherein the transparency information at least comprises transparency values of each pixel in each original image, performing color space transformation operation on each original image to obtain first images corresponding to each original image, wherein color information of each original image is represented through three primary color channels, color information in each first image is represented through a chromaticity channel and a brightness channel, performing image interception operation on each first image based on preset proportion to obtain second images corresponding to each first image, wherein each second image is a local image at the center position of each first image, performing block division on each second image to obtain M blocks corresponding to each second image, M is a positive integer, embedding the target coding sequence into the brightness channel corresponding to each block based on the embedding strength corresponding to each block in the M blocks to obtain M target blocks, and performing image interception operation on each first image to obtain second images