TW201015482A - System and method for removing watermarks from an image - Google Patents

Abstract

A system and method for removing watermarks from an image is provided. The method comprises: obtaining an original image from a storage of an electronic apparatus; decomposing the original image to a certain numbers of Eigen-images, and numbering each of the Eigen-images in sequence; identifying a number range of the Eigen-images that need to be removed; selecting all Eigen-images having watermarks from the number range according to the number of each of the Eigen-images; deleting the selected Eigen-images and combining the residual Eigen-images to form a combination image; checking whether the combination image is distorted; and storing the combination image if the combination image is not distorted.

Description

201015482 IX. INSTRUCTIONS: [Technical Field] The present invention relates to a digital image processing system and method, and more particularly to a system and method for removing watermark information from a digital image towel. [Prior Art] In the information age of the Internet, the speed of digital information transmission and control is quite fast. In order to prevent intellectual property rights of digital media from being harmed, especially for commercial images and music, it is currently Need to see the problem -. Therefore, many techniques for watermark information hiding have been widely proposed to protect or verify the owner's copyright information. With the rise of e-commerce] the business practices on the Internet are becoming more and more frequent, and the increase in Internet infringements has made digital media issues more and more important. For example, many image processing software have also added watermarks. Function to protect, the user's ~ * creation. However, since there is a technology to protect, there is of course a way to try to remove the watermark information. Seeing that the smear-printing attack is to shift the watermark information hidden in the image to the 'after the watermark information is removed, the image can still retain the original media, and the industry can successfully achieve the removal of the watermark information. Purpose. However: the shadow::::print removal technology can hide the watermark information in the ''bit is not guaranteed to be the quality of the image after the watermark information is removed' so that the image distortion after the watermark information is removed cannot be Preserve the commercial value of the original video media. [Summary of the Invention] Xin; X on each 'need to provide a watermark information removal system 201015482 卜 / / / / H shift m installation money in the electronic device for use in memory from the system includes: image access mode Group, 特征 之 特征 ( 四 成 成 成 成 成 成 ; 特征 特征 特征 特征 特征 特征 特征 特征 特征 特征 特征 特征 特征 特征 特征 特征 特征 特征 特征 特征 特征 特征 特征 特征 特征 特征 特征 特征 特征 特征 特征 特征 特征 特征 特征 特征 特征 特征 特征 特征 特征 特征 特征 特征, ",,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, The method of removing the information of the child's dream printing is applied to the electronic thank-you. The electric command 4 dares to recall the body. The method includes the following steps: (a) in the memory Ι ί: watermark image; (b) according to the floating water (four) image element The size 4 positive image is decomposed into a corresponding number of characteristic figures, and the: Zhang feature map is transferred to the wealth riding figure number; U) is decomposed into the feature = the graphic number range of the feature graphic to be removed is identified; (4) Yu Xin is out Μ形编号范_Screening (4) There are floating water (four) features of the second,] 4 selected feature graphics and the remaining feature graphics group and read (1) check whether the product f of the combined image is distorted; and (g) store the 4 distortion combined image to the memory of the memory 201015482 compared to the prior art, The watermarking information removal system and method can remove and remove the watermark information hidden in the watermark image by a singular value decomposition algorithm and a feature-image-identification distortion algorithm. The watermark information hidden in the watermark image can be removed, and the quality of the original image can be preserved. [Embodiment] Referring to FIG. 1, the watermark information removing system 1 of the present invention is preferably an embodiment.剌 ring (4). The material watermark information removal system is installed and operated on the electronic device towel, and the device includes a processing device 2 and a memory device 3. In the implementation, the electronic device rides the bicycle. ^Sub-devices can be used for computers, PDAs, and other processors with image processing functions. The processor 2 is used to run the watermark information removal system 2 to hide the material of the Koche image towel. : remember The volume 3 is used to store the watermarked image with watermark information and the combined image of the watermarked akizaki_. The watermark information removal system 1 is used to read the printed image from the memory 3, and the floating image The watermark image is decomposed into a certain number of characteristic graphics bands, :lmage) 'find and remove the feature map with watermark information'. The feature graphics of the 1j are combined into the image after the watermark information is removed: wide image and Checking the combined image quality and storing the combined image in f_3. In the actual towel, the watermark information is removed, and the image image access module 11, the image decomposition module is more than 12, and the feature graphic group is 1-port 3 feature graphic screening module U, watermarking information removal module 5 and image quality inspection module 16. 7 201015482

The image access module 11 is configured to read a floating image from the memory 3, and store the U water image obtained by removing the watermark information into the memory (4) 3. The watermark image is Hidden deer: Net watermark information 'For example, the copyright information of the protected object is hidden in the watermark image. Ο Image decomposition module, group 12 is used to decompose with a singular value (1 time Value DeC〇mposition' for short SVD The algorithm decomposes the watermark image into a fixed number of feature maps according to the size of the watermark==, and the number of each feature sheet is sequentially numbered. The shadow of the image ^8 solution module 12 can The watermark image is decomposed into 64 feature graphics, a knife feature graphic, or a larger number of feature graphics. In this embodiment, the original image size is a primitive, and the image decomposition module u decomposes the floating image into 128 features. Graphics, each feature graphic has a corresponding graphic number 'for example, feature graphic—the number is numbered as HU, 128 feature graphic. The SVD algorithm is a method to calculate the maximum of the watermark image by quantization. Singular value points to extract float The algorithm of printing information. Because the feature graphics are decomposed by SVD, the closer the feature image is to the original image, the closer it will be to the original image. Therefore, if 128 feature graphics are recombined, the combined image will be followed. If the original image is the same as the first 126 feature patterns, the difference between the combined image and the original image is indistinguishable to the user's naked eye. 'The feature graphics authentication module 13 is used to calculate the Distorable Distortion (JND) algorithm. The method is used to identify the range of graphic numbers to be removed. The jND calculus is used to calculate the characteristic size of the characteristic graphics (4) w e ® T_f. Whether the feature graphic contains the algorithm of the watermark information. Scope=Graphic teacher module 14 is used to calculate the feature value of all the feature points in the figure number = 2^, and add up all ^ = _ 特 魏 wei value, and according to the size of the total value of the feature screened with a feature graphic of floating water. If

The more the total value of a feature® shape is A, the hidden feature of the job profile is KP, but the smaller the total feature value of the feature image is, the watermark information hidden by the feature graphic is hidden. The less. The watermark-bein removal module 15 is used to delete the filtered feature patterns with floating print information and to combine the remaining __ shapes into a single image. The image quality checking module 16 is configured to calculate a Peak Signal to Noise Ratio (psNR value) of the combined image, and determine whether the quality of the combined image is based on whether the PSNR value exceeds a preset psNR allowable value. distortion. In this embodiment, the image quality inspection module 16 obtains the PSNR value of the combined image by calculating the ratio of the primitive value of the watermark image to the primitive value of the combined image. Generally, the user can set the PSNR allowable value of the watermark image according to the size of the primitive value of the watermark image. For example, the user can set the pSNR allowable value of the watermark image to 0.8. That is to say, if the psnr value of the combined image is greater than or equal to 0.8 ′, the distortion of the combined image relative to the watermark image is small, and the difference from the original image is indistinguishable from the user. Referring to Figure 2, a flow chart of a method for removing watermark information of the present invention is shown. In the embodiment, the watermarking information removing party 201015482 removes the watermarking information hidden in the watermark image by using the -SVD decomposition technology and the technology of the image of the bribe image. The watermark information hidden in the watermark image t can be removed, and the quality of the original image can be preserved. The image access module 11 reads a watermark image in the memory 3 (step S11). The floating water image hides the phase watermark information. For example, the watermark image hides the copyright information of the protection or verification owner. The image decomposing module 12 uses the SVD algorithm to decompose the watermark image into a number of feature patterns based on the image size of the watermark image and sequentially number each feature pattern (step S12). In this embodiment, the image of the county print is 128*128, and the image decomposition module 12 decomposes the watermark image into 128 feature graphics, and each of the feature graphics has a corresponding graphic number, for example, the feature graphic once. It is numbered as the U, 3...128 feature graphic. The feature о shape authentication module uses the nd algorithm to identify the feature number range of the feature graphic to be removed from the decomposed feature graphic (step su). In this embodiment, the feature graphic game module u will be I. : The number of graphics that need to be removed in the eight characteristic graphics ranges from the first to the first. How the feature pattern_selection module 13 utilizes the excitation algorithm for the feature number (4) that needs to be removed will be described in detail in FIG. The feature graphic screening module 14 selects the feature graphic having the watermark information within the range of the graphic number that has been played (step si4). The feature map 201015482 shape screening module 14 how to filter out having a watermark will be described in detail in FIG. The pattern of the watermark information removing module 15 deletes the filtered feature pattern of the message and the remaining feature pattern [: a combined image of the shell (step S15). After the attack is completed, the image quality inspection module 16 calculates the ratio of the image of the watermark image to the PSNr of the combined image to obtain the s(8). The image quality check module 16 determines the distortion of the combined image by determining the wake-up allowable value set by the job value step (step S17). Generally, the user can set the pSNR allowable value map of the watermark image according to the watermark no read size to set the PSNR allowable value of the watermark image to Q 8 . That is to say, the PSNR value of the two-combinable image is greater than (4)·8, which means that the combined image has a small degree of distortion of the watermark image, and the eyebrow and the eyebrow are relatively indistinguishable. The difference from the original shirt image is the user's naked eye. In step S17, if the quality of the combined image is distorted, that is, the PSNR allowable value set by the user of the combined image, then == return to step S13 to re-authenticate the need to move The range of the graphic number is removed to remove the watermark information in the hidden=watermark image. If the zero quality of the combined image is not distorted, that is, the psNR value of the combined image is greater than or equal to the user-defined PSNR allowable value. The image access module n stores the combined image without distortion in the memory 3 (step S18). Similarly, if the combined image is compared with the watermark distortion, the difference between the combined image and the original image is It is unacceptable to the naked eye and can be accepted by the user. 11 201015482 ^ As shown in Fig. 3, the detailed process method of step S13 in Fig. 2 shows how to use the jn diagram. The figure number range of the feature pattern that needs to be removed In the feature graphic ^ ^ 图形 馨 馨 馨 13 13 13 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 利用 J J J J J J J J J J J J J J J J J J J J J J J J J J JThe average JND value is calculated as the feature of all the pixel points in the chat circle image step S133). The brain value is shun a pixel point/do, the force is greater than the average JND佶Γ point of the JND value point ( Step S135). The 70 point of the figure is used as the characteristic feature of the feature graphic. The corresponding _ value in the graphic 13 module (step ^^ sign in the parent feature == the corresponding _ value in the feature graphic is close to The feature is programmed in the range ^ The graphic is placed in the graphic of the feature graphic to be removed = 2 7). In this embodiment, if there is a difference in the feature value in a certain feature to 49 feature graphics: The difference of the feature values in the 5th to 100th feature patterns is relatively close, and the difference in the feature values in the first to the 128th feature patterns is also larger than the pair. Therefore, the feature number of the feature pattern 13 is the range of the figure number to the third to the sheet. The feature graphic is placed in the range of the graphic number of the feature circle that needs to be removed. 12 201015482 . , See Figure 4, the system _2f refines the flow method, how to refine the flow chart. Characteristic graph of information. The following algorithm is screened to have a float-circle The characteristic figure "horse". The feature value of the feature feature graphic screening module 14 in the shape number field is represented as "step step ^ each element point ❹ 22 if the primitive value of the primitive point 仏 (4) Less than or equal to ^图__咖). The feature graphic (4) module 14 maps the feature map (4) $ ' == binary value "." (step S144 > if the primitive = average: the value of the five-one feature graphic _ = two of the primitive value _ will be = S145 ). On the feature map m, the value in the C step is M-selection, and the number of each of the -_signals is accumulated (step S146), and the number of feature points (step s147) is the total value of the feature, which is expressed as "and". The total value π to the small feature is sorted according to the large order (step S148). Finally, feature map 4 selects the feature map corresponding to the maximum feature total value buckle.

The symbol "圮" is deleted and the feature pattern & is removed within the range of the pattern number to be removed from the S' shape (step S149). In general, 13 201015482 If the feature points in the feature graph 5 r are (10), the larger the feature total value is, the more the watermark information is hidden in the feature graph. = _ is disclosed in the preferred embodiment as above, but it is not intended to limit = / Ming. Any modification of this invention may be made without departing from the spirit and scope of the invention, and the scope of the invention is defined by the scope of the appended claims. [Simple description of the map]

Q Figure 1 is an application environment diagram of a preferred embodiment of the watermarking information removal system of the present invention. Fig. 2 is a flow chart showing a preferred embodiment of the method for removing watermark information according to the present invention. FIG. 3 is a detailed flowchart of step S13 in FIG. 2. FIG. 4 is a detailed flowchart of step S14 in FIG. 2. [Main component symbol description] Watermark information removal system 1 Image access module 11 Image decomposition module 12 Feature graphics authentication module 13 Feature graphics screening module 14 Watermark information removal module 15 Image quality inspection module 16 Processor 2 Memory 3 Reads the watermark image from the memory. S11 201015482 ' Decomposes the watermark image into a feature graphic using the SVD algorithm. S12 uses the JND algorithm to identify the graphic number range S13. • Filters the feature graphic with watermark information. S14 • Delete the selected feature graphic and combine a combined image S15 to calculate the PSNR value of the combined image. S16 Determine whether the combined image is distorted according to the PSNR value. S17 Store the combined image into the memory. S18 〇❿ 15

Claims (1)

201015482 X. Patent Application Park 1. A watermark information removal system installed and operated on an electronic device. The electronic device includes a memory. The system includes: a centering, image access module for reading from a memory device A watermark I image decomposition module is used for decomposing the image of the primitive image of the watermark image into a corresponding number of feature graphics, and preparing a ', watermarking sequence _ New; # 征 graphics according to the characteristic graphic The identification module is used to extract the range of the graphic number from the extracted feature image; the moving feature graphic screening module is not required to calculate the feature of all the feature points in each feature graphic in the already-drawn graphic Value, the eigenvalue of the point of the ancient range to obtain the total value of the feature of the feature graphic, the size of the total value of the feature_outout the shape of the watermark information; and the feature watermarking information removal module for deleting (4) The selected feature patterns are combined into one combined image. Remaining φ 2·such as f please patent _ the first! The watermark information removal system of the item further includes an image quality checking module, configured to calculate the noise ratio of the composite image, and determine the peak of the image by the peak noise ratio (fiB' A-ji . 3. The watermarking information removal system of claim 2, wherein the image access module is further configured to store a group of people without distortion to the memory towel. 4. The watermarking information removal system of claim i, wherein the image decomposition model uses a singular value decomposition algorithm to decompose the floating water 16 201015482 * printed image into a feature graphic, the singular value The decomposition algorithm is an algorithm for extracting the watermark information hidden in the watermark image by calculating the maximum singular value point in the watermark image by the quantization strategy. The watermarking information removal system of claim 1, wherein the feature pattern discriminating module uses a discriminant distortion algorithm to identify a range of pattern numbers to be removed, the discriminant distortion calculation The law is an algorithm for determining whether the feature graphic contains watermark information by calculating the size of the feature value corresponding to the feature point in the feature graphic. 〇6·- A method for removing watermark information, which is applied to an electronic device, the electronic device comprising a memory, the method comprising the following steps: (a) reading a watermark image in the memory; (b) according to The size of the watermark image is decomposed into a corresponding number of feature graphics, and each feature graphic is sequentially numbered in sequence; (c) the feature graphic to be removed is identified in the extracted feature graphic (d) filtering the feature graphic with watermark information within the identified number of the graphic number; (e) deleting the selected feature graphic and combining the remaining feature graphics into a combined image (f) checking whether the quality of the combined image is distorted; and (g) storing the combined image without distortion in the memory. 7. The method for removing watermark information according to item 6 of the patent application scope, wherein the step (b) decomposes the watermark image into a corresponding number of feature patterns 17 201015482 ' is by a singular value decomposition algorithm In realization, the singular value decomposition algorithm is an algorithm for extracting the maximum singular value points in the watermark image by the quantization strategy, and extracting the watermark information hidden in the watermark image. -8. The method for removing watermark information as described in claim 6, wherein the step (c) comprises the steps of: converting the watermark image into a discriminating distortion image by using an identification distortion algorithm; Identifying the distortion value of each primitive point in the distorted image; ❹ calculating the average distortion value of all the primitive points in the discriminative distortion image; comparing the distortion value and the average distortion value of each primitive point; The feature point whose value is greater than the average distortion value is used as the feature point of the feature graphic; the corresponding feature value of the feature point in each feature pattern is compared; and the feature pattern whose feature point is close to the corresponding distortion value in the feature graphic is selected And put the feature graphic into the range of the graphic number of the feature graphic to be removed. ❿ 9. The method for removing watermark information as described in claim 6, wherein the step (d) comprises the steps of: calculating a primitive value of each primitive point in the feature graphic; and calculating all the primitives in the feature graphic The average element value of the point; determining whether the primitive value of each primitive point in the feature graphic is greater than the average primitive value; if the primitive value of the primitive point is less than or equal to the average primitive value, the map of the primitive point is The meta value is converted to the binary value "0", or, if the primitive value of the primitive point is greater than the 18 201015482 average primitive value, the primitive value of the primitive point is converted into binary == Zhang! The value is the number of feature points of "1", and each feature is purely selected according to the feature pattern corresponding to the maximum feature total value of the riding of the large material and the graphic dragon _ removed.胄The characteristic figure is 10·If the title of the application is 料6 ❹ ❹ The image of the combined image is opened ^4 “The value of the image and the value of the image; the ratio of the values gives the peak noise ratio of the combined image. ^The peak value of the combined image is the allowable value of the (4) household setting. If the peak noise ratio of the combined image is less than the allowable value set by the user, the quality of the combined image is already distorted; or if the combined image has a peak noise ratio A mosquito value greater than the equivalent of the household indicates that the quality of the combined image is not distorted.