TWI530914B - Data-bearing encoding system and decoding system utilizing halftone cell dot number modulation - Google Patents

Description

Halftone data steganography coding system and decoding system using block ink number modulation

The invention relates to a data encoding method, in particular to a halftone data steganography coding system and a decoding system for hiding data in a halftone pattern by modulating the number of ink dots in a block.

In order not to damage the overall beauty of the printed matter, the industry has been developing how to hide product information, such as anti-counterfeiting materials, in printed matter, such as trademark patterns or pictures, so that it can be integrated with printed content. For example, as shown in FIG. 1, U.S. Patent No. 8,854,453 discloses a method of encoding in a halftone image, which displays a dot shape of a halftone unit 502 composed of a plurality of pixels, and by moving the dots up and down. , left and right collapsed, so that the ink dot position can have four changes, so that halftone units 502, 504, 506, 508 having four different ink dot positions respectively represent a two-bit coded material, such as 00, 01, 10, 11, by which information can be steganized in halftone images in a two-bit encoding without affecting the overall aesthetics of the image. However, this method of coding deliberately moves the ink dots in a manner that is contrary to the traditional halftone conversion method and greatly changes the position of the ink dots. It is easy to make the quality of the original halftone image worse.

In addition, a direct binary search (hereinafter referred to as DBS) can optimize the quality of a halftone image, as shown in FIG. 2, that is, test each pixel in a halftone image. For example, after the pixel 10 located at the center of FIG. 2 is swapped by black and white or by swapping with the neighboring eight adjacent pixels, the halftone image and the original grayscale image are both passed through a human visual model (simulation). The visual perception of the human eye (ie, comparing the differences between the two) produces a perceptual error value, and continuously repeats the above DBS steps to recursively adjust the halftone image until the perceptual error value produced by the human visual model no longer changes. That is, the error between the halftone image and the original grayscale image has been minimized, and then the above action is repeated on the next pixel until all the pixels complete the DBS, thereby making the halftone image through the direct binary search Visually closest to the original grayscale image.

Therefore, how to reduce the influence on the halftone image quality in the halftone image encoding process, and to combine the DBS technology to optimize the quality of the encoded halftone image, has become the focus of the present invention.

Accordingly, it is an object of the present invention to provide a halftone data steganographic coding system and decoding system that utilizes the modulation of the number of dot ink dots.

Therefore, the present invention utilizes a halftone data steganographic coding system modulated by block ink dots to convert a grayscale image into a halftone image and steganography data. In the halftone image, the system includes: a halftone conversion module that converts the grayscale image into a halftone image having a plurality of halftone cells according to a mxm square threshold matrix, and each of the halftone cells Having mxm pixels, where m is a positive integer greater than 2; and an encoding module dividing the halftone units into a plurality of codeable blocks, each codeable block being adjacent to four halftones The unit is composed, and the encoding module includes 16 encoding rules corresponding to 16 kinds of encoded data, so as to select one of the encoding rules according to one of the encoding materials to modulate the number of ink dots of each halftone unit in the codeable block. And adding, subtracting 1 or maintaining the number of ink dots of each halftone unit in the codeable block, so that the codeable block becomes an encoded block implied by the coded data, and the half is made The tonal image becomes a coded halftone image.

In an embodiment of the invention, the halftone data steganography system further includes an image optimization module that performs a direct binary search on the halftone units in the encoded halftone image, and once Performing a direct binary search for only one halftone unit, the direct binary search determines that one of the halftone cells is located in the middle, corner or side of the halftone unit, and attempts to apply the ink dot The positions of the non-ink dot pixels adjacent thereto are interchanged until it is determined that one of the interchange results causes the halftone unit to be closest to a portion of the image corresponding to the gray-scale image, that is, the result of the interchange is accepted, and the next ink dot is repeatedly executed. The direct binary search, until all the ink dots complete the direct binary search, the image optimization module performs the direct binary search for the next halftone unit, and the encoded halftone image is The halftones in The unit performs the direct binary search one by one from top to bottom and left to right, thereby generating an optimized encoded halftone image.

In addition, the present invention is used to decode the encoded halftone image, and a halftone data steganography decoding system using block dot number modulation is used to read data hidden in the encoded halftone image, and The method includes: a database that records an original grayscale image corresponding to the encoded halftone image and a mxm square threshold matrix, the position of the encoded block in the encoded halftone image, and the corresponding 16 encoded data. 16 coding rules; a half tone conversion module, which converts the original grayscale image into a halftone image according to the mxm square threshold matrix; and a decoding module according to the coded region in the encoded halftone image Positioning the block to find at least one coded block in the encoded halftone image and comparing the ink of each halftone unit in the coded block corresponding to the coded block and the halftone image a number of points to find an encoding rule for converting the codeable block into the coded block from the 16 coding rules, and obtaining a code corresponding to the coding rule Information.

In an embodiment of the invention, the mxm square threshold matrix comprises a screen vector consisting of two two-dimensional vectors [m, 0], [0, m] and mxm threshold values.

In an embodiment of the invention, the threshold values are generated by a centralized dithering method or a decentralized dithering method.

The effect of the invention is: by placing four adjacent images in a halftone image The halftone unit constitutes one codeable block, and according to the 16 encoding rules corresponding to the 16 kinds of coded data, the number of dots of each halftone unit in the codeable block is incremented, decremented or maintained. The 16 kinds of coded data are steganized in the coded block, so that the coded block becomes an encoded block with implicit four-bit data, and the ink in the halftone image is not greatly changed during the encoding process. The position of the dot does not affect the image quality of the encoded halftone image, and the image optimization module performs a pair of coded halftone images in a halftone unit in a top-to-bottom, left-to-right manner. Each of the halftone units performs a direct binary search, which not only optimizes the encoded halftone image, but also reduces the number of direct binary search executions, thereby increasing the speed of image processing.

1‧‧‧ halftone conversion module

2‧‧‧Code Module

3‧‧‧Image to be encoded

3’‧‧‧ Coded imagery

4‧‧‧Image Optimization Module

30‧‧‧ Grayscale imagery

31‧‧‧ halftone image

32‧‧‧Coded image

33‧‧‧Optimized coded imagery

40‧‧‧square threshold matrix

51‧‧‧Database

52‧‧‧ halftone conversion module

53‧‧‧Decoding module

301‧‧‧ Block

311‧‧‧ halftone unit

312‧‧‧codeable block

313‧‧‧ Coded Blocks

Other features and effects of the present invention will be apparent from the following description of the drawings, wherein: FIG. 1 is a schematic diagram of a conventional halftone image encoding method for encoding by moving a dot shape; A schematic diagram of a direct binary search method is shown; FIG. 3 is a schematic diagram showing the components of an embodiment of the halftone data steganography system using the dot number modulation of the present invention; FIG. 4 illustrates the halftone of the embodiment. The conversion module converts grayscale images into halftone images; FIG. 5 illustrates that the four adjacent halftone units in the halftone conversion module 1 form one codeable block in this embodiment; FIG. 6 shows 16 encoding rules corresponding to 16 types of encoded data in this embodiment; FIG. The coding module of the embodiment writes an encoded data into a coded block; FIG. 8 illustrates that the encoded halftone image of the embodiment is processed by the image optimization module to become an optimized coded halftone image. FIG. 9 illustrates that the image optimization module of the present embodiment first determines that the ink dots are located in the middle, corner, or side of the halftone unit when performing direct binary search; FIG. 10 shows that the present invention utilizes the number of dot ink dots. Schematic diagram of constituent elements of an embodiment of a variable halftone data steganographic decoding system; and FIG. 11 illustrates that the database in FIG. 10 records a square threshold matrix that converts the original grayscale image into a halftone image, and corresponding ten Sixteen encoding rules for six kinds of encoded data, and the decoding module compares the halftone image with the encoded halftone image to obtain a corresponding encoding rule, and obtains an encoding corresponding to the encoding rule. Code data.

Before the present invention is described in detail, it should be noted that in the following description, similar elements are denoted by the same reference numerals.

Referring to FIG. 3, an embodiment of a halftone data steganography coding system using the patch number of blocks in the present invention is used to convert a grayscale image into a halftone image. Like, and sneak data in the halftone image. The system includes a halftone conversion module 1 and an encoding module 2. The grayscale image 30 is an example of an image having 16×16 pixels, which represents a small portion of a to-be-coded image 3. The halftone conversion module 1 converts the grayscale image 30 into a halftone image 31 according to a square threshold matrix 40.

Halftone is a method commonly used in the printing industry to process tone and simulate continuous tone. The images printed on the printer can only be represented by the two tones of ink or no ink. The binary image like this is called halftone image. Therefore, by adjusting the arrangement of the ink dots on the printing plane, the principle of "visual integration" of the human eye can be utilized, so that the halftone image reproduces the feeling of continuously adjusting the image under a certain distance observation. So halftone technology is a technique for converting continuously tuned images into halftone images. The halftone screen method (also known as ordered dithering method) first divides a continuous tone image (grayscale image) into a plurality of blocks of the same size as a threshold matrix, and then sets their grayscale values. Comparing with a critical value of the relative position in the threshold matrix, if the grayscale value of the relative position is greater than the critical value, the white point is output (no ink), and vice versa, the black point (inking) is output, thereby continuously converting the image A halftone image.

When the halftone screen method is used, the arrangement and shape of the dots of the generated halftone image are affected by the order of the critical values in the critical matrix. Therefore, the conventional threshold matrix can be mainly divided into the clustered-dot Ordered Dithering method and the scattered dithering method. (Dispersed-Dot Ordered Dithering) two, as shown in the following two figures.

In addition, when using the halftone screen method, there are two factors that affect the threshold matrix. The first one is the order of the critical values in the matrix, such as the aforementioned "concentrated dithering method" and "distributed dithering". Color method." The second is the Screen Tile Vector, which determines the shape and rotation angle of the threshold matrix. The screen vector contains two two-dimensional vectors n ₁ , n ₂ , and the two two-dimensional vectors n ₁ , n ₂ determine the number of pixels in each halftone cell in the halftone image (below Will be further explained). Since the "threshold matrix" is a conventional technique applied to the traditional halftone screen method, and is not the focus of this discussion, it will not be described in detail here. For related technologies, see "Design of Color Screen Tile Vector Sets" and "Computer-aided design of clustered-dot color screens based on a human visual system model" and other papers.

Therefore, in the present embodiment, as shown in FIG. 4, the square threshold matrix 40 employed is a 4x4 distributed threshold matrix, which is composed of two two-dimensional vectors [4, 0], [0, 4] A screen vector composed of a matrix consisting of 4x4 thresholds. Of course, this embodiment can also adopt a centralized threshold matrix.

Therefore, as shown in FIG. 4, the halftone conversion module 1 divides the grayscale image 30 into 16 blocks 301 in units of 4×4 pixels according to the size of the square threshold matrix 40, and then from top to bottom. From left to right, the grayscale value of the 4x4 pixels of each block 301 is compared with the critical value of the relative position in the threshold matrix 4, and if the grayscale value of the pixel is greater than the critical value, the output 1 represents white, otherwise The output 0 represents black (ink dot), whereby the grayscale image 30 is converted into a halftone image 31 displayed only in black and white, the halftone image 31 includes a plurality of halftone units 311, and each halftone unit 311 Has 4x4 pixels.

Next, the halftone image 31 is input to the encoding module 2, and as shown in FIG. 5, the encoding module 2 first divides the halftone units 311 of the halftone image 31 into a plurality of codeable blocks 312, and In this embodiment, as shown in FIG. 5, each codeable block 312 is composed of four adjacent halftone units 311. Of course, the codeable block may also be composed of four consecutive halftone units 311 that are laterally or longitudinally adjacent.

As shown in FIG. 6, the encoding module 2 includes 16 encoding rules corresponding to 16 kinds of encoding data 0000, 0001, 0010, 0011, . . . 1111. For example, the first encoding rule corresponding to the encoded data 0000 is The number of dots of each halftone unit 311 in the encoding block 312 remains unchanged (indicated by 0 in the figure), and the second encoding rule corresponding to the encoded material 0001 is a halftone of the upper left of the codeable block 312. The number of ink dots of the unit 311 is increased by 1 (indicated by +1 in the figure), and the number of ink dots of the halftone unit in the upper right of the codeable block 312 is decremented by 1 (indicated by -1 in the figure), and the remaining halftone units are 311 ink dots The number remains unchanged (indicated by 0 in the figure), and the third encoding rule corresponding to the encoded data 0010 is to decrement the number of dots of the halftone unit 311 in the upper left of the codeable block 312 by 1 and to make the codeable area The number of dots of the halftone unit 311 in the upper right of the block 312 is incremented by 1, and the number of dots of the remaining halftone cells 311 remains unchanged, and the rest of the encoding rules are deduced by analogy.

And the coded data is expressed in binary, for example, four bits 0000 represent a decimal number 0, four bits 0011 represent a decimal number 3, and the like. Thereby, when one of the encoded materials, for example, 0001, is to be loaded into the halftone image 31, a second encoding rule corresponding to the encoded material 0001 is selected to modulate one of the codeable blocks in the halftone image 31. The number of dots of each halftone unit 311 in 312. And generally, the coding direction is from top to bottom, from left to right. Therefore, the present embodiment takes the first (first upper left) codeable block 312 of the halftone image 31 into the first coded data 0001 as an example. . Since the second encoding rule corresponding to the encoded material 0001 is to increase the number of dots in the halftone unit 311 in the upper left of the codeable block 312 by 1, the number of dots in the upper right halftone unit 311 is reduced. 1, the number of dots in the halftone unit 311 in the lower left and lower right remains unchanged. Therefore, as shown in FIG. 7, the number of ink dots of the upper left, upper right, lower left, and lower right four halftone units 311 in the codeable block 312 is originally 8, 9, 8, and 9, respectively, and the encoded The number of ink dots of the upper left, upper right, lower left, and lower right four halftone units 311 in the encoding block 313 will become 9, 8, 8, and 9, respectively.

Moreover, when to be added in the halftone unit 311 of the codeable block 312 When an ink dot is used, the ink dot may be placed at a pixel position of any non-ink dot (ie, white dot) in the halftone unit 311, and when one dot in the halftone cell 311 is to be reduced, halftone may be Any ink dot in cell 311 is replaced with a non-ink dot (ie, a white dot).

Therefore, according to the above encoding rule, the four-bit encoded data can be loaded into the codeable block 312 by modulating the number of dots of each halftone unit 311 in the codeable block 312. An encoded block 313 of the encoded material is hidden, and the halftone image having the encoded block 313 is made into a coded halftone image 32, as shown in FIG. The encoding module 2 records the position of the encoded block 313 in the encoded halftone image 32 in a database (not shown) for later use in decoding.

Therefore, each codeable block 312 of this embodiment can hide (carry) four-bit coded data. And since the four-bit data is loaded in each of the codeable blocks 312, the encoding rule of the present embodiment only adds at most one dot or one dot to each halftone unit 311, for example, when the encoded data 1111 is loaded. At the very least, it is not necessary to change the number of ink dots in each halftone unit 311 at all, for example, when the encoded material 0000 is loaded, and most of the encoding rules only need to change two and a half of each of the codeable blocks 312. The number of dots of the tone unit 311 (the number of dots is increased or decreased by one), so that the dot position in the halftone image is not greatly changed during the encoding process, and the image quality of the encoded halftone image is not affected.

Furthermore, in order to make the encoded halftone image 32 have better image quality As shown in FIG. 8 , the embodiment further includes an image optimization module 4 for performing a direct binary search (hereinafter referred to as DBS) on the halftone units 311 in the encoded halftone image 32. That is, the image optimization module 4 performs DBS on only one halftone unit 311 at a time. For example, as shown in FIG. 9, the direct binary search first determines that one dot in the halftone unit 311 is located in the halftone unit 311. In the middle (as shown by 9A in Fig. 9), the corner (as shown by 9B in Fig. 9) or the side (as shown by 9C in Fig. 9), and the ink dot is only around it (as shown in Fig. 9). In the dotted line frame of 9A, 9B, and 9C, one of the adjacent non-ink dots (ie, white dots) is interchanged, and the halftone unit 311 after the pixel is interchanged is corresponding to the partial image in the original grayscale image. The two sense (ie, compare the difference between the two) through a human visual system filter kernel to generate a perceptual error value, and continuously repeat the above steps, that is, the ink dots are respectively Another non-ink point (ie, white point) pixel adjacent to each other is interchanged and continuously passed through the human eye vision model. The perceptual error value is generated until it is found that one of the interchange results minimizes the perceptual error value and does not change, and the human visual model determines that the error of the halftone unit 311 and the corresponding partial image in the original grayscale image has been minimized, that is, It is determined that the halftone unit 311 is closest to the corresponding partial image in the original grayscale image, and the result of the interchange is accepted. Then, DBS is repeatedly executed on the next ink dot in the halftone unit 311 until all the ink dots in the halftone unit 311 complete the DBS, and then the DBS is performed on the next halftone unit 311, thereby generating An encoded halftone image 33 is optimized. Because the above DBS is a well-known technology, and it is not the focus of this case, Without further elaboration, related technologies can be found in the paper "A Novel Bareode System for Intelligent Automation Industry".

The image optimization module 4 also records the position of the encoded block 313 in the optimized encoded halftone image 33 in a database (not shown) for later use in decoding.

Since the direct binary search (DBS) performed by the image optimization module 4 of the present embodiment causes only the ink dots in the halftone unit 311 to change its position (SWAP) in the halftone unit 311, the encoded region The number of dots of the halftone unit 311 in block 313 can remain unchanged while retaining the encoded material implied therein.

In addition, since the image optimization module 4 of the present embodiment performs a direct binary search by using each of the halftone units 311 in the pair of coded halftone images 32 from top to bottom and left to right. Therefore, the image optimization module 4 performs a direct binary search on a halftone unit 311 until the error of the halftone unit 311 and the corresponding partial image in the original grayscale image converges to a minimum. The next halftone unit 311 performs a direct binary search, thereby reducing the number of direct binary search executions and increasing the speed of image processing.

Therefore, when a pattern or trademark is presented as a coded halftone image, such as the optimized coded halftone image 33 described above, in order to read the data steganized in the optimized halftone image 33 for subsequent verification or application As shown in FIG. 10, an embodiment of a halftone data steganography decoding system using the modulation of the dot number of blocks in the present invention For example, the data used in the optimized coded halftone image 33 is extracted, and mainly includes a database 51, a halftone conversion module 52, and a decoding module 53. As shown in Fig. 11, the optimized encoded halftone image 33 represents a portion of the image of an encoded image 3'. The database 51 mainly records the original grayscale image 3 (represented by the partial grayscale image 30 in the figure) corresponding to the encoded image 3' (indicated by the optimized encoded halftone image 33 in the figure) for using the grayscale image. 30 converts the square threshold matrix 40 of the halftone image 31, optimizes the position of the encoded block 313 in the encoded halftone image 33, and the above 16 encoding rules corresponding to the 16 encoded data.

Therefore, as shown in FIG. 10 and FIG. 11, when the decoding module 53 receives the optimized halftone image 33, it notifies the halftone conversion module 52, and the halftone conversion module 52 is taken out from the database 51. The original grayscale image 30 corresponding to the optimized coded halftone image 33, and the original grayscale image 30 is converted into the halftone image 31 according to the square threshold matrix 40, as in the foregoing halftone conversion module 1. Provided to the decoding module 53.

Then, as shown in FIG. 11, the decoding module 53 finds out the optimized encoded halftone image 33 based on the position of the encoded block in the optimized encoded halftone image 33 recorded in the database 51. Encoding block 313 and comparing the number of dots of the four halftone cells in the encoded block 313 with the four halftone cells corresponding to the halftone image 31, for example, as shown in FIG. 11, the coded block The number of ink dots (9) in the upper left halftone unit in 313 is larger than that in the halftone unit 311 corresponding to the halftone image 31 The number of ink dots (8) is one more, and the number of ink dots (8) in the upper right halftone unit in the encoded block 313 is smaller than the number of ink dots (8) in the corresponding halftone unit 311 in the halftone image 31. 1, and the number of ink dots in the halftone unit at the lower left and the lower right of the encoded block 313 is the same as the number of ink dots in the corresponding halftone unit 311 in the halftone image 31, so the decoding module 53 can According to the difference in the number of ink dots between the above-mentioned encoded block 313 and the corresponding four halftone units in the halftone image 33, a corresponding one encoding rule (ie, the second encoding rule) is found from the 16 encoding rules. And obtaining an encoded data 0001 corresponding to the encoding rule, and outputting the decoding result. Accordingly, the present embodiment can read the material hidden in the encoded halftone image 3' in accordance with the above decoding method.

In summary, the present invention composes the codeable region by composing four adjacent halftone units 311 in the halftone image 31 into one codeable block 312 and according to 16 coding rules corresponding to the 16 coded materials. The number of dots of each halftone unit 311 in block 312 is incremented, decremented, or remains unchanged, and the 16 types of encoded data are steganized in the codeable block 312, making the codeable block 312 an implicit The encoded block 313 of the four-bit data does not significantly change the position of the ink dot in the halftone image during the encoding process, without affecting the image quality of the encoded halftone image 32, and is optimized by the image. Group 4 performs a direct binary search for each halftone unit 311 in a pair of encoded halftone images 32 in a halftone unit, in a top-to-bottom, left-to-right manner, except that the second half of the encoding can be performed. Optimized tonal image 32, and can reduce the number of direct binary search execution, and improve the speed of image processing, and indeed achieve this The efficacy and purpose of the invention.

However, the above is only the embodiment of the present invention, and the scope of the invention is not limited thereto, and all the equivalent equivalent changes and modifications according to the scope of the patent application and the patent specification of the present invention are still The scope of the invention is covered.

1‧‧‧ halftone conversion module

2‧‧‧Code Module

4‧‧‧Image Optimization Module

Claims (7)

A halftone data steganography coding system utilizing block ink dot modulation for converting a grayscale image into a halftone image and rewriting data in the halftone image; the system comprises: halftone conversion a module that converts the grayscale image into a halftone image having a plurality of halftone cells according to a mxm square threshold matrix, and each of the halftone cells has mxm pixels, where m is a positive integer greater than two; And an encoding module, the halftone unit is divided into a plurality of codeable blocks, each codeable block is composed of four adjacent halftone units, and the coding module includes corresponding 16 types of coded data. 16 encoding rules for modulating the number of dots of each halftone unit in the codeable block according to one of the encoding parameters, and modulating the number of dots of each halftone unit in the codeable block The number of dots is incremented by one, decremented by one or remains unchanged, so that the codeable block becomes an encoded block that implies the encoded material, and the halftone image becomes a coded halftone image. The halftone data steganography system for modifying the number of ink dots in a block according to claim 1, further comprising an image optimization module for directly performing the halftone units in the encoded halftone image Two-bit search, and perform direct binary search for only one halftone unit at a time, the direct binary search determines that one dot in the halftone unit is located in the middle, corner or side of the halftone unit And attempting to swap the ink dot with its adjacent non-ink dot pixel position until it is determined that one of the interchange results causes the halftone cell to be closest to a portion of the image corresponding to the grayscale image, that is, According to the result of the exchange, the direct binary search is repeated for the next ink dot until all the ink dots complete the direct binary search, and the image optimization module performs the next halftone unit again. The direct binary search, and the direct binary search is performed one by one from top to bottom and left to right for the halftone units in the encoded halftone image, thereby generating an optimized coding Rear halftone image. A halftone data steganography system using a block ink dot modulation as claimed in claim 1 or 2, wherein the mxm square critical value matrix comprises one by two [m, 0], [0, m] A vector of vectors consisting of dimensional vectors and mxm thresholds. A halftone data steganography system using a block ink dot modulation as described in claim 3, wherein the threshold values are generated by a centralized dithering method or a decentralized dithering method. A halftone data steganography decoding system for utilizing block ink dot modulation for reading data steganized in a coded halftone image, and comprising: a database whose record corresponds to the encoded halftone image An original grayscale image and a mxm square threshold matrix, the position of the encoded block in the encoded halftone image, and 16 encoding rules corresponding to the 16 encoded data; and a halftone conversion module according to the mxm a square threshold matrix for converting the original grayscale image into a halftone image; and a decoding module for finding at least one of the encoded halftone images based on the position of the encoded block in the encoded halftone image Coded a block, and comparing the number of ink dots of each halftone unit in the coded block corresponding to the coded block and the halftone image, to find out that the coded code can be encoded from the 16 coding rules The block is converted into an encoding rule of the coded block, and an encoded data corresponding to the encoding rule is obtained. A halftone data steganography decoding system using a block ink dot modulation as described in claim 5, wherein the mxm square critical value matrix comprises two two-dimensional vectors of [m, 0], [0, m] A mesh vector constructed and mxm thresholds. The halftone data steganography decoding system using the block ink dot modulation as described in claim 6, wherein the threshold values are generated by a centralized dithering method or a decentralized dithering method.