An anti-counterfeiting authentication method based on statistical differences of bone width transformation

Abstract

The present invention provides an anti-counterfeiting authentication method based on the statistical difference of bone width transformation, which uses bone width transformation to extract the key area of ​​the image to be tested, and realizes the function of quickly identifying the authenticity of the anti-counterfeiting image to be tested by comparing the connectivity of the key areas of the sample image and the image to be tested. The present invention improves the performance of the anti-counterfeiting query and authentication system based on image technology, improves the robustness of identification, reduces the misjudgment rate, and provides a reliable basis for further distinguishing the authenticity of the anti-counterfeiting pattern; the one-object-one-code method is adopted, which increases the cost and difficulty of counterfeiting. The operation method of the present invention is simple and easy, and there is no need to learn identification knowledge and use specific equipment. The identification operation can be performed by taking pictures with a mobile phone carried with you. It is efficient and convenient, improves the identification speed, and has broad application prospects.

Description

Technical Field

[0001] The invention belongs to the technical field of anti-counterfeiting and information security, and in particular relates to an anti-counterfeiting authentication method based on statistical differences of bone width transformation. Background Art

[0002] For a long time, counterfeit and shoddy products have caused serious harm to the world economy and consumers. Product anti-counterfeiting has always been a concern for people, among which anti-counterfeiting labels are the most commonly used anti-counterfeiting means. Common anti-counterfeiting labels refer to labels that can be pasted, printed, or transferred on the surface of a product or packaging and have an anti-counterfeiting effect. Traditional methods include intuitive methods, hand touch methods, texture methods, perspective methods, and instrument detection methods, but these methods are either not very reliable or inconvenient to operate. With the rapid development of smart phones, it is becoming easier and easier to obtain digital images. Using digital image technology to identify the authenticity of anti-counterfeiting labels has become a fast, convenient, and reliable anti-counterfeiting method. It is also a development trend of current product anti-counterfeiting technology. Therefore, the research on methods and systems for automatic authenticity recognition of anti-counterfeiting patterns based on digital images has become an urgent requirement for social and economic development.

[0003] However, due to the complex background environment in practice, the existing image-based artificial intelligence technology used in anti-counterfeiting authentication has the problem of low recognition accuracy, which makes the application of image technology in the field of printing anti-counterfeiting more challenging. At present, anti-counterfeiting methods do not use machine vision identification very much, and mainly use anti-counterfeiting materials, which is not convenient for inexperienced users to verify. Summary of the invention

[0004] The technical problem to be solved by the present invention is to provide an anti-counterfeiting authentication method based on the statistical difference of bone width transformation, which is used to quickly identify the authenticity of an anti-counterfeiting image to be tested.

[0005] The technical solution adopted by the present invention to solve the above technical problems is: an anti-counterfeiting authentication method based on the statistical difference of bone width transformation, comprising the following steps:

[0006] S1: Randomly generate a unique anti-counterfeiting pattern with a fine strip texture and bind it to a unique product; store the anti-counterfeiting pattern in the cloud authentication platform as a sample image;

[0007] S2: The user collects the anti-counterfeiting pattern of the product and uploads it to the cloud authentication platform as the image to be tested;

[0008] S3: The cloud authentication platform matches the identity code of the image to be tested with the identity code of the preset sample image, and corrects the image to be tested using the sample image;

[0009] S4: performing binarization processing on the sample image and the corrected image to be tested respectively, extracting the skeleton of the binary image to obtain a skeleton image, and fusing the skeleton image and the binary image to obtain a fused image;

[0010] S5: using a bone width transformation algorithm to extract the key area of ​​the image to be tested and the key area of ​​the sample image;

[0011] S6: Identify the authenticity of the image to be tested by comparing the difference between the key area of ​​the image to be tested and the key area of ​​the sample image.

[0012] According to the above scheme, in step S1, the specific steps are:

[0013] S11: Randomly generate a unique anti-counterfeiting pattern with a fine strip texture;

[0014] S12: Based on the randomness of ink flow, a printer is used to print an anti-counterfeiting pattern to generate an anti-counterfeiting mark and bind it to the displaced product, so that the product and the anti-counterfeiting pattern have one object and one code, and one-to-one correspondence;

[0015] S13: Collect the digital image of the anti-counterfeiting mark and store it in the cloud authentication platform as a sample image.

[0016] According to the above scheme, in step S5, the specific steps are:

[0017] S51: Perform bone width transformation calculation on the fusion image; calculate the normal direction θ of each point on the skeleton according to the skeleton function f(x) on the fusion image of the sample image and the fusion image of the corrected image to be tested, then the normal direction of point a on the skeleton is expressed as:

[0018]

[0019] Assume that the pixel value at position (i, j) in the skeleton binary image is g(i, j), perform gradient calculation on the skeleton to obtain the gradient change values ​​on both sides of the skeleton line, and use the horizontal and vertical gradients to obtain the gradient direction τ:

[0020]

[0021]

[0022]

[0023] The gradient direction of the skeleton is regarded as the normal direction, and the edge points are searched according to the inner and outer normal directions respectively. When the background point is found, the traversal is stopped, and this point is the edge point sought:

[0024]

[0025] set up and is the pixel value of the position (i, j) in the binary image of the anti-counterfeiting image after moving along the inner and outer directions of τ, then the distance between the two edge points q and p constitutes the skeleton width w = ||qp||;

[0026] S52: Remove outliers through denoising, remove endpoints and single points, and take the remaining skeleton points that are smaller than the skeleton width threshold t as key points; assuming the skeleton width of the skeleton point is d(·), the unit step function sq(·) for determining the key point is:

[0027]

[0028] S53: Select an n×n template to perform connected domain analysis, connect the skeleton points to determine the size of the connected area; then construct a rectangle for each connected area and expand it by m pixels around to form a key area.

[0029] Furthermore, in the step S52, the bone width threshold t is set to 3 pixels, and the image is normalized to 400*400; in the step S53, n=5, m=2.

[0030] Furthermore, in step S52, the specific steps are: when a sudden change occurs in the bone width value of a certain point, the bone width of the point is treated as a noise point and removed through denoising.

[0031] Furthermore, in step S52, the specific steps are as follows: when the bone width value of a point is less than the bone width threshold t, the point is regarded as a subtle area; if the point is the only point within the h×h window that is less than the bone width threshold t, the point is considered to be a single point or endpoint and is removed.

[0032] Furthermore, in step S6, the specific steps are:

[0033] S61: corresponding all key areas to the connected domains of the binary image of the sample image and the connected domains of the binary image of the image to be tested, respectively, and performing 8-neighborhood connected domain analysis on the binary image of the image to be tested and the binary image of the sample image;

[0034] S62: Counting the connected labels of each key area, comparing the difference between the key area of ​​the image to be tested and the key area of ​​the sample image, and determining the authenticity of the image to be tested;

[0035] Let N d _boundary(i) and N s _boundary(i) is the number of connected labels of the image to be tested and the sample image in the i-th subtle region, and the difference judgment formula of each key region is:

[0036]

[0037] S63: If the sum of the number of differences in the subtle areas of the image to be tested Sum(Num(i)) is greater than the judgment threshold ρ, the image to be tested is judged to be counterfeit; if the sum of the number of differences in the subtle areas of the image to be tested Sum(Num(i)) is less than or equal to the judgment threshold ρ, the image to be tested is judged to be authentic:

[0038]

[0039] Furthermore, in step S63, the judgment threshold ρ is set to 30% of the total number of subtle regions.

[0040] A computer storage medium stores a computer program executable by a computer processor, wherein the computer program executes an anti-counterfeiting authentication method based on statistical differences in bone width transformation.

[0041] The beneficial effects of the present invention are:

[0042] 1. The present invention discloses an anti-counterfeiting authentication method based on statistical difference of bone width transformation, which uses bone width transformation to extract key areas of the image to be tested, and realizes the function of quickly identifying the authenticity of the anti-counterfeiting image to be tested by comparing the connectivity of the key areas of the sample image and the image to be tested.

[0043] 2. The present invention improves the performance of the anti-counterfeiting query and authentication system based on image technology, improves the robustness of identification, reduces the misjudgment rate, and provides a reliable basis for further distinguishing the authenticity of the anti-counterfeiting pattern; the one-object-one-code approach is adopted, which increases the cost and difficulty of counterfeiting.

[0044] 3. The operation method of the present invention is simple and easy, and there is no need to learn identification knowledge or use specific equipment. The identification operation can be performed by taking photos with a mobile phone carried with you. It is efficient and convenient, improves the identification speed, and has broad application prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 is a flow chart of an embodiment of the present invention.

[0046] Figure 2 It is an anti-counterfeiting pattern with random fine textures according to an embodiment of the present invention.

[0047] Figure 3 It is a diagram showing the skeleton extraction processing effect of the anti-counterfeiting pattern according to an embodiment of the present invention.

[0048] Figure 4 This is a diagram showing the processing effect of bone width transformation according to an embodiment of the present invention.

[0049] Figure 5 It is a noise map of calculation abnormality in the bone width transformation in the embodiment of the present invention.

[0050] Figure 6 4 is a diagram of a fine region construction process according to an embodiment of the present invention.

[0051] Figure 7 It is a fine area diagram of the anti-counterfeiting pattern according to an embodiment of the present invention.

[0052] Figure 8 It is a connected domain labeling graph of the anti-counterfeiting pattern binary graph according to an embodiment of the present invention.

[0053] Fig. 9 It is a connectivity difference map of the sample image and the forged image in the key area of ​​the embodiment of the present invention. DETAILED DESCRIPTION

[0054] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0055] The technical solution adopted by the present invention is: first, different anti-counterfeiting labels with texture features are pasted or printed on the product, and the image is collected and uploaded to the cloud-based anti-counterfeiting original information database as a sample; then, the customer takes a photo of the anti-counterfeiting label to be tested on the product and uploads it to the certification center, the skeleton width transformation algorithm is used to extract the key area of ​​the sample, and the differences in the key areas of the anti-counterfeiting image to be tested are compared to identify its authenticity.

[0056] See also Figure 1 , a method according to an embodiment of the present invention comprises the following steps:

[0057] S1: Paste or print an anti-counterfeiting pattern with random small strips on the product, such as Figure 2 As shown. Each anti-counterfeiting pattern is different and unique, ensuring that each product has a unique code. At the same time, the characteristic information of the anti-counterfeiting pattern will be entered into the cloud-based anti-counterfeiting original information database in advance as a sample for subsequent anti-counterfeiting authentication procedures.

[0058] S2: The customer uses dedicated software to capture images of anti-counterfeiting logos on the products being sold, and the software automatically uploads the images to the authentication center on the cloud server for verification.

[0059] S3: On the cloud server, the identity code of the image to be tested is matched with the identity code of the original sample image on the cloud, and then the sample image is used to correct the image to be tested.

[0060] S4: Binarize the sample image and the corrected image to be tested, extract the skeleton of the binary image, fuse the skeleton image and the binary image, and calculate the bone width transformation, such as Figure 3 As shown in the figure, after the skeleton of the white area is extracted, it is fused into the binary image, and the thin lines in the fused image represent the skeleton.

[0061] The fused image first obtains the normal direction θ of each point on the skeleton. The normal direction can be calculated based on the skeleton function f(x). The normal direction of point a on the skeleton can be expressed as:

[0062]

[0063] The skeleton in the image exists in the form of a discrete point set. Its skeleton function needs to be obtained by curve fitting with the point set, and then the normal direction of each point is calculated. This process is complex and time-consuming. Therefore, the gradient of the skeleton is calculated to obtain the gradient change values ​​on both sides of the skeleton line. The gradient calculation can use a simple horizontal and vertical difference method to obtain the gradient direction τ using the horizontal and vertical gradients. The calculation formula is as follows:

[0064]

[0065]

[0066]

[0067] Where g(i,j) is the pixel value at position (i,j) in the skeleton binary image. The gradient direction of the skeleton can be regarded as its normal direction. Then, edge points are searched according to the directions of the inner and outer normals respectively. When a background point is found, the traversal stops, and this point is the required edge point.

[0068]

[0069] in and is the pixel value of (i, j) in the binary image of the anti-counterfeiting image after moving along the inner and outer directions of τ. The distance between the two edge points q and p constitutes the skeleton width w = ||qp||, as Figure 4 As shown, the polyline is the extracted skeleton, points p and q are pixels on the bone boundary, the arrow is the normal direction, and w is the bone width.

[0070] S5: Bone width transformation provides the width of the pattern texture in the binary image of the anti-counterfeiting pattern. However, due to calculation deviations at some corners during skeleton extraction, the calculation of bone width is abnormal, such as Figure 5 However, these abnormal values ​​have a low frequency of occurrence and can be treated as noise points. That is, when the bone width value changes dramatically, the bone width at this point is a noise point and needs to be denoised.

[0071] S6: In order to extract the key area of ​​the anti-counterfeiting pattern, that is, the strokes with narrow width, the position with smaller width is found according to the bone width value. For example, when the bone width value is less than the width threshold t, the point will be regarded as a subtle area. Since simply using the width threshold as the basis for judgment will cause some endpoints and single points to be misjudged as subtle areas, these interference points need to be removed. Single points can be judged according to the number of points in the connected domain. For example, if the point is the only one within the 3×3 window that is less than the width threshold, it is considered to be a single point or endpoint and is removed.

[0072] S7: After outlier denoising and elimination of endpoints and single points, the remaining skeleton points that are smaller than the bone width threshold are used as key points. The unit step function sq(·) for determining key points is as follows:

[0073]

[0074] Where d(·) is the bone width of the skeleton point, and t is the bone width threshold, which is generally set to 3 pixels (the image is normalized to 400*400).

[0075] S8: For the construction of key areas, we cannot simply construct the area with a single skeleton key point, especially for some narrow and long subtle areas. The adjacent skeleton key points that meet the conditions can be connected together to form a large key area. Therefore, a 5×5 template is selected for connected domain analysis, the skeleton points are connected, the size of the connected area is determined, and then a rectangle is constructed for each connected area and expanded by 2 pixels around to form a key area, such as Figure 7 shown.

[0076] S9: In order to identify the authenticity of the anti-counterfeiting pattern to be tested, it is necessary to perform 8-neighborhood connected domain analysis on the binary image of the anti-counterfeiting pattern to be tested and the binary image of the reserved sample. Then, the connected domain analysis of the key areas of the anti-counterfeiting pattern to be tested and the reserved sample is performed to determine the authenticity of the anti-counterfeiting pattern to be tested. The connected domain analysis is to correspond all the key areas to the connected domains of the sample binary image and the binary image of the anti-counterfeiting pattern to be tested, and count the connected labels of each key area, and make comparative judgments. Usually, the narrow and thin strokes in the anti-counterfeiting pattern are disconnected during counterfeit printing, generally forming two connected domains, and the small gaps between the strokes are filled during counterfeit printing, and the two connected domains will become one connected domain, such as Fig. 9 As shown, it is obvious that the connected areas contained in the red box of the sample image of the anti-counterfeiting pattern (Figure a) and the forged image to be tested (Figure b) are different. The difference judgment formula of each key area is:

[0077]

[0078] Where N d _boundary(i) and N s_boundary(i) is the number of connected labels that the anti-counterfeiting pattern to be tested and the reserved samples have in the i-th subtle area.

[0079] S10: When the number of differences in the subtle areas of the anti-counterfeiting pattern to be tested exceeds a certain threshold, it can be considered to be a forged anti-counterfeiting pattern:

[0080]

[0081] The total number of differences is represented by Sum(Num(i)), and ρ is a judgment threshold, which is generally set to 30% of the total number of subtle areas. If it is less than the threshold, the anti-counterfeiting pattern to be tested is considered to be authentic.

[0082] The above embodiments are only used to illustrate the design ideas and features of the present invention, and their purpose is to enable those skilled in the art to understand the content of the present invention and implement it accordingly. The protection scope of the present invention is not limited to the above embodiments. Therefore, any equivalent changes or modifications made based on the principles and design ideas disclosed by the present invention are within the protection scope of the present invention.

Claims

1. An anti-counterfeiting authentication method based on the statistical difference of bone width transformation, characterized by: The following steps are involved: S1: Randomly generate a unique anti-counterfeiting pattern with a fine strip texture and bind it to a unique product; store the anti-counterfeiting pattern in the cloud authentication platform as a sample image; S2: The user collects the anti-counterfeiting pattern of the product and uploads it to the cloud authentication platform as the image to be tested; S3: The cloud authentication platform matches the identity code of the image to be tested with the identity code of the preset sample image, and corrects the image to be tested using the sample image; S4: performing binarization processing on the sample image and the corrected image to be tested respectively, extracting the skeleton of the binary image to obtain a skeleton image, and fusing the skeleton image and the binary image to obtain a fused image; S5: using the bone width transformation algorithm to extract the key area of ​​the image to be tested and the key area of ​​the sample image; the specific steps are: S51: Perform bone width transformation calculation on the fusion image; calculate the normal direction θ of each point on the skeleton according to the skeleton function f(x) on the fusion image of the sample image and the fusion image of the corrected image to be tested, then the normal direction of point a on the skeleton is expressed as: Assume that the pixel value at position (i, j) in the skeleton binary image is g(i, j), perform gradient calculation on the skeleton to obtain the gradient change values ​​on both sides of the skeleton line, and use the horizontal and vertical gradients to obtain the gradient direction τ: The gradient direction of the skeleton is regarded as the normal direction, and the edge points are searched according to the inner and outer normal directions respectively. When the background point is found, the traversal is stopped, and this point is the edge point sought: set up and is the pixel value of the position (i, j) in the binary image of the anti-counterfeiting image after moving along the inner and outer directions of τ, then the distance between the two edge points q and p constitutes the skeleton width w = ||qp||; S52: Remove outliers through denoising, remove endpoints and single points, and take the remaining skeleton points that are smaller than the skeleton width threshold t as key points; assuming the skeleton width of the skeleton point is d(·), the unit step function sq(·) for determining the key point is: S53: Select an n×n template to perform connected domain analysis, connect the skeleton points to determine the size of the connected area; then construct a rectangle for each connected area and expand it by m pixels around to form a key area; S6: Identify the authenticity of the image to be tested by comparing the difference between the key area of ​​the image to be tested and the key area of ​​the sample image.

2. The anti-counterfeiting authentication method based on the statistical difference of bone width transformation according to claim 1 is characterized in that: In the step S1, the specific steps are: S11: Randomly generate a unique anti-counterfeiting pattern with a fine strip texture; S12: Based on the randomness of ink flow, a printer is used to print an anti-counterfeiting pattern to generate an anti-counterfeiting mark and bind it to the displaced product, so that the product and the anti-counterfeiting pattern have one object and one code, and one-to-one correspondence; S13: Collect the digital image of the anti-counterfeiting mark and store it in the cloud authentication platform as a sample image.

3. The anti-counterfeiting authentication method based on the statistical difference of bone width transformation according to claim 1 is characterized in that: In the step S52, the bone width threshold t is set to 3 pixels, and the image is normalized to 400*400; In the step S53, n=5, m=2.

4. The anti-counterfeiting authentication method based on the statistical difference of bone width transformation according to claim 1 is characterized in that: In the step S52, the specific steps are: When the bone width value of a certain point changes suddenly, the bone width of this point is treated as a noise point and removed through denoising.

5. The anti-counterfeiting authentication method based on the statistical difference of bone width transformation according to claim 1 is characterized in that: In the step S52, the specific steps are: When the bone width value of a point is less than the bone width threshold t, the point is regarded as a subtle area; if the point is the only point within the h×h window that is less than the bone width threshold t, the point is considered to be a single point or endpoint and is removed.

6. The anti-counterfeiting authentication method based on the statistical difference of bone width transformation according to claim 5 is characterized in that: In the step S6, the specific steps are: S61: corresponding all key areas to the connected domains of the binary image of the sample image and the connected domains of the binary image of the image to be tested, respectively, and performing 8-neighborhood connected domain analysis on the binary image of the image to be tested and the binary image of the sample image; S62: Counting the connected labels of each key area, comparing the difference between the key area of ​​the image to be tested and the key area of ​​the sample image, and determining the authenticity of the image to be tested; Let N d _boundary(i) and N s _boundary(i) is the number of connected labels of the image to be tested and the sample image in the i-th subtle region, and the difference judgment formula of each key region is: S63: If the sum of the number of differences in the subtle areas of the image to be tested Sum(Num(i)) is greater than the judgment threshold ρ, the image to be tested is judged to be counterfeit; if the sum of the number of differences in the subtle areas of the image to be tested Sum(Num(i)) is less than or equal to the judgment threshold ρ, the image to be tested is judged to be authentic:

7. The anti-counterfeiting authentication method based on the statistical difference of bone width transformation according to claim 6 is characterized in that: In the step S63, The judgment threshold ρ is set to 30% of the total number of fine areas.

8. A computer storage medium, characterized in that: A computer program executable by a computer processor is stored therein, and the computer program executes an anti-counterfeiting authentication method based on statistical differences of bone width transformation as described in any one of claims 1 to 7.

Images