CN103310236A - Mosaic image detection method and system based on local two-dimensional characteristics - Google Patents
Mosaic image detection method and system based on local two-dimensional characteristics Download PDFInfo
- Publication number
- CN103310236A CN103310236A CN2013102616210A CN201310261621A CN103310236A CN 103310236 A CN103310236 A CN 103310236A CN 2013102616210 A CN2013102616210 A CN 2013102616210A CN 201310261621 A CN201310261621 A CN 201310261621A CN 103310236 A CN103310236 A CN 103310236A
- Authority
- CN
- China
- Prior art keywords
- local
- absolute value
- histogram
- matrix
- dimensional
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Abstract
The invention belongs to the technical field of image processing and information security and relates to a mosaic image detection method and system based on local two-dimensional characteristics. The mosaic image detection method includes that images are cut through squares with different side lengths and then are subjected to blocking Discrete Cosine Transformation (DCT), and the obtained blocked DCT coefficients are described in a local two-dimensional characteristic mode, are merged to be an integral detection characteristic and then are classified by a classifier. The mosaic image detection method and system can consider both the detection accuracy and the detection complexity, and the detection accuracy can reach 89.9%.
Description
Technical field
What the present invention relates to is that a kind of image is processed and the method and system of field of information security technology, specifically a kind of for stitching image detection method and system based on local two dimensional character to the stitching image that do not have priori.
Background technology
The society Digital image technology is very universal, and the threshold of use also reduces increasingly.A lot of powerful digital imaging processing softwares are also contacted by masses, and the ordinary people through professional training can not produce the forgery image that naked eyes can't directly be distinguished yet.When the forgery image comes across in the focus models such as forum, microblogging, cause great negative effect can for government, enterprise or individual, therefore forge the study hotspot that picture becomes current information content safety field for adopting computer technology to detect.
Usually, the forgery image detect technology of main flow is divided into two large classes at present: active mode and passive mode.Active mode mainly is to embed digital signature or digital watermarking etc. in the process of generating pictures, guarantees that by the detection to these marks picture is not tampered.Passive mode is then mainly utilized the statistical property of picture itself, and does not rely on the identifiable marker of prior implantation.The marks such as the digital signature that active mode is implanted and digital watermarking have certain destructiveness for picture itself, should not use in some cases.The adaptability of opposite passive mode is stronger, therefore becomes current main direction of studying.
The splicing of image is the most basic step of forging image.Complete distorted image flow process has generally included splicing, convergent-divergent, rotation and subsequent treatment, is the basis of most false proof authentication methods for the detection of concatenation.
Classical Image Mosaics detection method has the bispectrum feature detection method of the propositions such as Ng, see Ng TT, Chang SF., the data set of the true and stitching image piece of A dataset of authentic and spliced image blocks(), (ADVENT Technical Report, #203-2004-3, Columbia University.) document provides a general splicing image detect data set simultaneously, in order to the quality of more various algorithms, extensively quoted.Ng etc. have obtained 72% Detection accuracy at this data set.The people such as Fu adopt Hilbert-Huang transform and the moment characteristics of the fundamental function of wavelet transformed domain to detect, and have obtained 80.15% Detection accuracy.
Different with above-mentioned statistical method is, the people such as Johnson utilize the inconsistency of the different splicing regions illumination of stitching image feature to detect, see Johnson MK, Farid H.Exposing digital forgeries by detecting inconsistencies in lighting (forging identification based on the numeral that the illumination inconsistency detects) .(In Proceedings of ACM Multimedia and Security Workshop, New York, USA, 2005; 1 – 9.) employing, comprises the steps such as feature selecting, feature extraction and classification learning, thereby becomes detection method commonly used because have ripe method frame based on the detection method of statistical nature.But existing detection method accuracy in detection based on statistical nature is still waiting to improve, and also has simultaneously a large amount of statistical natures not to be used to the detection of this problem.
Find through the retrieval to prior art, Chinese patent literature CN102855496A, open day 2013-01-02 discloses a kind of face authentication method and system of blocking, and this technology comprises: S1, collection people face video image; S2, the people's face video image that gathers is carried out pre-service; S3, carry out detection computations to blocking people's face, according to the movable information of video sequence, utilize three frame difference methods that the position of facial image is estimated, then carry out the affirmation of further people's face position by the Adaboost algorithm; S4, identify calculating to blocking people's face, people's face sample is divided into some piecemeals, adopt the two minutes algorithms of SVM in conjunction with supervision 1-NN nearest neighbour method that people's face piecemeal is blocked differentiation, if piecemeal is blocked, then directly give up, if piecemeal is not blocked, then extract corresponding LBP texture feature vector and be weighted identification, then use the sorter based on the rectangular projection method to be used for reducing the characteristic matching number of times.This technology mechanically is divided into 6 zones with image, although can be with solving people's face and face's relatively-stationary problem in vitals position, the problem of identifying without any the passive image forge of priori for image but can't solve.
Summary of the invention
The present invention is directed to the prior art above shortcomings, propose a kind of stitching image detection method and system based on local two dimensional character, can take into account accuracy of detection and detection complexity.
Described local two dimensional character refers to a kind of local dual mode (Local Binary Pattern that detects customization for splicing, LBP) feature, the LBP feature is people's propositions such as Ojala the earliest, see Ojala T, Pietikainen M, Maenpaa T.Multiresolution gray-scale and rotation invariant texture classification with local binary patterns.(adopts gray scale and the invariable rotary Texture classification of the multiresolution of local dual mode) (IEEE Transactions on Pattern Analysis and Machine Intelligence2002; 24 (7): 971 – 987.)
Described LBP feature is widely used in recognition of face, background modeling and steganalysis as a kind of image texture characteristic.The core concept of LBP is that an image pixel and the pixel around him are compared, and after then processing through thresholding, obtains 0/1 sequence.0/1 sequence can be regarded a binary number as, thereby represents with a metric integer.The corresponding decimal integer of each pixel in the image array.These integers can consist of a histogram, and this histogram has reflected the statistical laws such as edge of image.
The present invention is achieved by the following technical solutions:
The present invention relates to a kind of stitching image detection method based on local two dimensional character, carry out block DCT transform after cutting apart by the square that image is adopted the different length of sides, and after the mode of the local two dimensional character of the piecemeal DCT coefficients by using that obtains is described and merges into complete detected characteristics, adopt sorter to classify.
Described method specifically may further comprise the steps:
Step 1: treat arbitrary image that deal with data concentrates and carry out multiple dimensioned piecemeal DCT(discrete cosine) conversion, the piecemeal DCT matrix of coefficients that obtains, and piecemeal DCT coefficient all taken absolute value, obtain piecemeal DCT coefficient absolute value matrix.
Described block DCT transform refers to: select a kind of side length b, pending image is divided into the square tiles that the length of side is the formed objects of b, then carry out dct transform in each square tiles.
The side length b of described square tiles is preferably 8 multiple, as be generally 8,16,32 ..., but also can make any number according to application demand.
As the big or small aliquant b of pending image, then add 0 in the rightmost side and the lower side of pending image, until it meets the integral multiple of b; Resulting piecemeal DCT matrix of coefficients is the matrix of a M * N behind the process block DCT transform, and pending picture size is that the two is equirotal to M ' * N ' usually.When added 0 row and 0 row in the pending image rightmost side and lower side after, resulting matrix of coefficients can be greater than pending image.
Step 2: adopt local two dimensional character to come token image to splice the variation of the statistical nature that causes, piecemeal DCT coefficient absolute value matrix is converted into local two dimensional character histogram.
2.1 on piecemeal DCT coefficient absolute value matrix arbitrary element around get P point, be denoted as g
p, p={1 ... P}.These points are minute arrangements such as 2 π/P angle around central point.When peripheral point not on matrix grid point, utilize its peripheral net point numerical value to carry out interpolation estimation, the distance of peripheral point and central point is designated as R.
P preferred value of some upper and lower, left and right that are described arbitrary element and upper left, lower-left, upper right, bottom right totally 8 points in the present invention, the distance R of peripheral point and central point preferably gets 1.
2.2 the gray-scale value of peripheral point is compared with the gray-scale value of central point successively: when the gray-scale value of the peripheral point gray-scale value greater than central point, then comparative result is denoted as 1, otherwise is 0; Then the comparative result of P point is arranged according to right-to-left, consisted of 0/1 comparison result sequence that length is P, this comparison result sequence also is used for being converted to decimal integer as bigit.
2.3 each element on the piecemeal DCT coefficient absolute value matrix is carried out the processing of step 2.2, obtain corresponding decimal integer, all decimal integers are consisted of a local two dimensional character histogram.
Step 3: the length of side value b that gets different square tiles in the step 1 is the corresponding absolute value matrix that obtains representing the piecemeal DCT coefficient of various yardsticks then, and obtains the local two dimensional character histogram of several correspondences according to the operation of step 2; The feature that each local two dimensional character histogram is generated is together in series and consists of a complete statistical nature, then utilizes the svm classifier device to learn and classify.
The block DCT transform of getting different edge long value b in the described local two dimensional character histogram in each unique corresponding step 1.
Described histogrammic horizontal scale value is 0 to 2
P-1, histogrammic value corresponding to each horizontal scale as a feature, characteristic dimension 2
P
That described sorter adopts is the SVM realization LibSVM of main flow, concrete such as Chang CC, Lin CJ.LIBSVM:a library for support vector machines.http: //www.csie.ntu.edu.tw/cjlin/libsvm, 2001.
3.1 obtain the piecemeal DCT coefficient absolute value matrix of different scale by getting different length of side value b, each piecemeal DCT absolute value matrix adopted step 2 is described to obtain local two dimensional character histogram.
3.2 extract 2 from each local two dimensional character histogram
PDimensional feature, and these features are serially connected form a complete statistical nature.
3.3 all the other all pictures of pending data centralization are extracted feature according to above-mentioned feature extracting method, simultaneously picture is divided into training set and test set two parts, first feature and the categorical data of training set are inputted sorter and obtained disaggregated model; Again the feature of disaggregated model and test set is inputted sorter again, the classification that obtains test set is differentiated; Last classification according to known test set obtains classify accuracy.
The number of pictures ratio of described training set and test set is preferably 5:1;
The splicing picture that comprises in described training set and the test set and the number of natural picture preferably are 1:1.
The present invention relates to a kind of stitching image detection system based on local two dimensional character, comprise: pretreatment module, local two-dimensional histogram makes up module, characteristic extracting module and classifier modules, wherein: pretreatment module links to each other with local two-dimensional histogram structure module and receives original image and carries out obtaining piecemeal DCT coefficient absolute value matrix and export local two-dimensional histogram to making up module behind the block DCT transform, local two-dimensional histogram structure module links to each other with characteristic extracting module and piecemeal DCT coefficient absolute value matrix is carried out obtaining local two-dimensional histogram and outputing to characteristic extracting module after the LBP computing, characteristic extracting module is connected with classifier modules and local two-dimensional histogram information is carried out obtaining characteristic of division after the feature extraction computing and outputs to classifier modules, and classifier modules receives characteristic of division and carries out obtaining the classification of original image is judged after the sort operation.
Technique effect
Compared with prior art, the present invention effectively catches the flaw that the splicing picture causes by adopting multiple dimensioned block DCT transform.Be determined by experiment the parameter of optimum local two dimensional character algorithm, P=8, R=1.The present invention has higher Detection accuracy than existing technology.
Description of drawings
Fig. 1 is schematic flow sheet of the present invention.
Fig. 2 is the parameter declaration schematic diagram of LBP algorithm among the embodiment.
Fig. 3 is system architecture schematic diagram of the present invention.
Fig. 4 is embodiment treatment effect schematic diagram;
Among the figure: (1) is that former figure, (2) are that LBP histogram, (3) under 8 * 8 block DCT transforms are that LBP histogram, (4) under 16 * 16 block DCT transforms are the LBP histogram under 32 * 32 block DCT transforms; Because 0 numerical exception is large, can affect histogrammic demonstration, all omitted 0 value here.
Embodiment
The below elaborates to embodiments of the invention, and the present embodiment is implemented under take technical solution of the present invention as prerequisite, provided detailed embodiment and concrete operating process, but protection scope of the present invention is not limited to following embodiment.
As shown in Figure 1, the present embodiment may further comprise the steps:
Three kinds of macroblock modes among the Fig. 1 that in Fig. 1, provides in (2), 4x4,8x8 and 16x16 have both guaranteed accuracy in detection through experimental results show that, have again less characteristic dimension.
Table 1 is LBP
8,1Accuracy in detection under different piecemeal DCT patterns, accuracy numerical value are the mean value of the testing result of 20 random sample groupings, and the numerical value in the bracket is square error.The picture library that adopts is that the stitching image of Columbia University detects the storehouse, see Ng TT, Chang SF., the data set of the true and stitching image piece of A dataset of authentic and spliced image blocks(), (ADVENT Technical Report, #203-2004-3, Columbia University.)
Table 1
The pixel pixel number relatively that is used on every side is denoted as P.These pixels are minute arrangements such as 2 π/P angle around central point.When neighboring pixel point not on image lattice point, adopt the interpolation representation of its peripheral net point numerical value.Another variable is R, and this variable represents is the distance of compared pixels point and central point on every side.Because the value of each representative of binary number is 2
p, so binary sequence is converted to decimal numeral formula and can be expressed as:
The combination of P and R has (8,1) usually, (8,2) and (8,3).Except these parameter combinations of single consideration, also they can be combined forms larger feature, i.e. described multiscale analysis.Result in the table 1 is the result of pattern (8,1), below further provide the result of multiscale analysis in the table 2.
Table 2 is the multiscale analysis result
| (P,R) | (8,1)+(8,2) | (8,1)+(8,2)+(8,3) |
| Accuracy | 90.45% | 90.48% |
Through comparative experiments, balance accuracy in detection and characteristic dimension determine to adopt single yardstick P=8, and R=1 is best LBP descriptor parameter.The s function by getting different σ, between 0 to 2, take 0.1 as step-length, observes the variation tendency that accuracy in detection is fallen after rising.When σ=0.9, accuracy in detection is the highest.The parameter that adopt (4) among Fig. 1 as mentioned above.
In LibSVM, need to determine the kernel function that adopt, through alternative Gauss RBF kernel function.Gauss RBF kernel function has two variablees: punishment variable C and gaussian kernel width gamma.
Adopt the mode of grid search to determine best C and γ combination, wherein: the hunting zone of C is 2
{ 1,1,3,5}, the hunting zone of γ is 2
5 ,-3 ,-1,1}
The parameter of above-mentioned C and γ is to consisting of a grid.Carry out cross validation at training set and find best parameter pair, then utilize this parameter to carrying out final test.Result in above-mentioned table 1 and the table 2 adopts said method to obtain.
As shown in Figure 3, for realizing the system of said method, this system comprises: pretreatment module, local two-dimensional histogram makes up module, characteristic extracting module and classifier modules, wherein: pretreatment module links to each other with local two-dimensional histogram structure module and receives original image and carries out obtaining piecemeal DCT coefficient absolute value matrix and export local two-dimensional histogram to making up module behind the block DCT transform, local two-dimensional histogram structure module links to each other with characteristic extracting module and piecemeal DCT coefficient absolute value matrix is carried out obtaining local two-dimensional histogram and outputing to characteristic extracting module after the LBP computing, characteristic extracting module is connected with classifier modules and local two-dimensional histogram information is carried out obtaining characteristic of division after the feature extraction computing and outputs to classifier modules, and classifier modules receives characteristic of division and carries out obtaining the classification of original image is judged after the sort operation.
What table 3 showed is on the stitching image detection storehouse of Columbia University, the comparison of the Detection accuracy of the detection method of usefulness the present invention and existing two kinds of main flows.Be respectively the Markov algorithm of Shi, Shi, Yun Q., Chunhua Chen, and Wen Chen. " A natural image model approach to splicing detection. " (a kind of natural picture model for detection of the splicing picture) Proceedings of the9th workshop on Multimedia﹠amp; Security.ACM, 2007, bispectrum algorithm with Ng, Ng, Tian-Tsong, Shih-Fu Chang, and Qibin Sun. " Blind detection of photomontage using higher order statistics. " (based on the splicing picture blind Detecting of high-order statistic) Circuits and Systems, 2004.ISCAS'04.Proceedings of the2004International Symposium on.Vol.5.IEEE, 2004.
Table 3
| ? | The present invention | Markov | Two spectrums |
| Accuracy rate | 89.9% | 86.6% | 72.3% |
Upper table shows that algorithm of the present invention has better Detection accuracy than existing technology.
Claims (10)
1. the stitching image detection method based on local two dimensional character is characterized in that, may further comprise the steps:
Step 1: treat arbitrary image that deal with data concentrates and carry out multiple dimensioned block DCT transform, the piecemeal DCT matrix of coefficients that obtains, and piecemeal DCT coefficient all taken absolute value, obtain piecemeal DCT coefficient absolute value matrix;
Step 2: adopt local two dimensional character to come token image to splice the variation of the statistical nature that causes, piecemeal DCT coefficient absolute value matrix is converted into local two dimensional character histogram;
Step 3: the length of side value of getting different square tiles in the step 1 is the corresponding absolute value matrix that obtains representing the piecemeal DCT coefficient of various yardsticks then, and obtains the local two dimensional character histogram of several correspondences according to the operation of step 2; The feature that each local two dimensional character histogram is generated is together in series and consists of a complete statistical nature, then utilizes the svm classifier device to learn and classify.
2. method according to claim 1 is characterized in that, described block DCT transform refers to: select a kind of side length b, pending image is divided into the square tiles that the length of side is the formed objects of b, then carry out dct transform in each square tiles.
3. method according to claim 2 is characterized in that, the side length b of described square tiles is 8 multiple; As the big or small aliquant b of pending image, then add 0 in the rightmost side and the lower side of pending image, until it meets the integral multiple of b; Resulting piecemeal DCT matrix of coefficients is the matrix of a M * N behind the process block DCT transform, and pending picture size is that the two is equirotal to M ' * N ' usually; When added 0 row and 0 row in the pending image rightmost side and lower side after, resulting matrix of coefficients can be greater than pending image.
4. method according to claim 1 is characterized in that, described step 2 comprises following operation:
2.1 on piecemeal DCT coefficient absolute value matrix arbitrary element around get P point, be denoted as g
p, p={1 ... P}; These points are minute arrangements such as 2 π/P angle around central point; When peripheral point not on matrix grid point, utilize its peripheral net point numerical value to carry out interpolation estimation, the distance of peripheral point and central point is designated as R;
2.2 the gray-scale value of peripheral point is compared with the gray-scale value of central point successively: when the gray-scale value of the peripheral point gray-scale value greater than central point, then comparative result is denoted as 1, otherwise is 0; Then the comparative result of P point is arranged according to right-to-left, consisted of 0/1 comparison result sequence that length is P, this comparison result sequence also is used for being converted to decimal integer as bigit;
2.3 each element on the piecemeal DCT coefficient absolute value matrix is carried out the processing of step 2.2, obtain corresponding decimal integer, all decimal integers are consisted of a local two dimensional character histogram.
5. method according to claim 4, it is characterized in that, described on piecemeal DCT coefficient absolute value matrix arbitrary element around get P and put and to refer to: in the upper and lower, left and right of described arbitrary element and upper left, lower-left, upper right, bottom right totally 8 points, the distance R of peripheral point and central point gets 1.
6. method according to claim 4 is characterized in that, described histogrammic horizontal scale value is 0 to 2
P-1, histogrammic value corresponding to each horizontal scale as a feature, characteristic dimension 2
P
7. method according to claim 1 is characterized in that, described step 3 comprises following operation:
3.1 obtain the piecemeal DCT coefficient absolute value matrix of different scale by getting different length of side value b, each piecemeal DCT absolute value matrix adopted step 2 is described to obtain local two dimensional character histogram;
3.2 extract 2 from each local two dimensional character histogram
PDimensional feature, and these features are serially connected form a complete statistical nature;
3.3 all the other all pictures of pending data centralization are extracted feature according to above-mentioned feature extracting method, simultaneously picture is divided into training set and test set two parts, first feature and the categorical data of training set are inputted sorter and obtained disaggregated model; Again the feature of disaggregated model and test set is inputted sorter again, the classification that obtains test set is differentiated; Last classification according to known test set obtains classify accuracy.
8. method according to claim 7 is characterized in that, the number of pictures ratio of described training set and test set is 5:1.
9. method according to claim 7 is characterized in that, the splicing picture that comprises in described training set and the test set and the number of natural picture are 1:1.
10. detection system that is used for realizing the described method of above-mentioned arbitrary claim, it is characterized in that, comprise: pretreatment module, local two-dimensional histogram makes up module, characteristic extracting module and classifier modules, wherein: pretreatment module links to each other with local two-dimensional histogram structure module and receives original image and carries out obtaining piecemeal DCT coefficient absolute value matrix and export local two-dimensional histogram to making up module behind the block DCT transform, local two-dimensional histogram structure module links to each other with characteristic extracting module and piecemeal DCT coefficient absolute value matrix is carried out obtaining local two-dimensional histogram and outputing to characteristic extracting module after the LBP computing, characteristic extracting module is connected with classifier modules and local two-dimensional histogram information is carried out obtaining characteristic of division after the feature extraction computing and outputs to classifier modules, and classifier modules receives characteristic of division and carries out obtaining the classification of original image is judged after the sort operation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2013102616210A CN103310236A (en) | 2013-06-27 | 2013-06-27 | Mosaic image detection method and system based on local two-dimensional characteristics |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2013102616210A CN103310236A (en) | 2013-06-27 | 2013-06-27 | Mosaic image detection method and system based on local two-dimensional characteristics |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN103310236A true CN103310236A (en) | 2013-09-18 |
Family
ID=49135430
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2013102616210A Pending CN103310236A (en) | 2013-06-27 | 2013-06-27 | Mosaic image detection method and system based on local two-dimensional characteristics |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103310236A (en) |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103544703A (en) * | 2013-10-19 | 2014-01-29 | 侯俊 | Digital image stitching detecting method |
| CN103914839A (en) * | 2014-03-27 | 2014-07-09 | 中山大学 | Image stitching and tampering detection method and device based on steganalysis |
| CN103955691A (en) * | 2014-05-08 | 2014-07-30 | 中南大学 | Multi-resolution LBP textural feature extracting method |
| CN103996044A (en) * | 2014-05-29 | 2014-08-20 | 天津航天中为数据系统科技有限公司 | Method and device for extracting targets through remote sensing image |
| CN104244016A (en) * | 2014-08-12 | 2014-12-24 | 中山大学 | H264 video content tampering detection method |
| CN104598929A (en) * | 2015-02-03 | 2015-05-06 | 南京邮电大学 | HOG (Histograms of Oriented Gradients) type quick feature extracting method |
| CN104899846A (en) * | 2015-05-20 | 2015-09-09 | 上海交通大学 | Digital image splicing passive detection method based on frequency domain local statistic model |
| CN106056523A (en) * | 2016-05-20 | 2016-10-26 | 南京航空航天大学 | Digital image stitching tampering blind detection method |
| CN106203492A (en) * | 2016-06-30 | 2016-12-07 | 中国科学院计算技术研究所 | The system and method that a kind of image latent writing is analyzed |
| CN108520215A (en) * | 2018-03-28 | 2018-09-11 | 电子科技大学 | Single sample face recognition method based on multiple dimensioned union feature encoder |
| CN108645875A (en) * | 2018-03-20 | 2018-10-12 | 上海市建筑科学研究院 | A kind of defect identification method of precast shear wall grouting connection |
| CN109086801A (en) * | 2018-07-06 | 2018-12-25 | 湖北工业大学 | A kind of image classification method based on improvement LBP feature extraction |
| CN109885987A (en) * | 2019-01-24 | 2019-06-14 | 中山大学 | A kind of bianry image steganalysis method based on directionality local binary patterns |
| CN111145146A (en) * | 2019-12-11 | 2020-05-12 | 北京航空航天大学 | Method and device for detecting style migration forged image based on HHT |
| CN111415336A (en) * | 2020-03-12 | 2020-07-14 | 泰康保险集团股份有限公司 | Image tampering identification method and device, server and storage medium |
| CN112232162A (en) * | 2020-10-06 | 2021-01-15 | 武汉烽火凯卓科技有限公司 | Pedestrian detection method and device based on multi-feature fusion cascade classifier |
| CN113837976A (en) * | 2021-09-17 | 2021-12-24 | 重庆邮电大学 | Multi-focus image fusion method based on combined multi-domain |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080193031A1 (en) * | 2007-02-09 | 2008-08-14 | New Jersey Institute Of Technology | Method and apparatus for a natural image model based approach to image/splicing/tampering detection |
| US20110002504A1 (en) * | 2006-05-05 | 2011-01-06 | New Jersey Institute Of Technology | System and/or method for image tamper detection |
| US20110019907A1 (en) * | 2006-01-13 | 2011-01-27 | New Jersey Institute Of Technology | Method for identifying marked images using statistical moments based at least in part on a jpeg array |
-
2013
- 2013-06-27 CN CN2013102616210A patent/CN103310236A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110019907A1 (en) * | 2006-01-13 | 2011-01-27 | New Jersey Institute Of Technology | Method for identifying marked images using statistical moments based at least in part on a jpeg array |
| US20110002504A1 (en) * | 2006-05-05 | 2011-01-06 | New Jersey Institute Of Technology | System and/or method for image tamper detection |
| US20080193031A1 (en) * | 2007-02-09 | 2008-08-14 | New Jersey Institute Of Technology | Method and apparatus for a natural image model based approach to image/splicing/tampering detection |
Non-Patent Citations (2)
| Title |
|---|
| YUJIN ZHANG等: "Image-splicing forgery detection based on local binary patterns of DCT coefficients", 《WILEY ONLINE LIBRARY》, 26 February 2013 (2013-02-26), pages 2386 - 2395 * |
| 张震等: "一种新的拼接图像检测方法", 《计算机应用研究》, vol. 26, no. 3, 15 March 2009 (2009-03-15), pages 1127 - 1130 * |
Cited By (26)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103544703B (en) * | 2013-10-19 | 2016-12-07 | 上海理工大学 | Digital picture splicing detection method |
| CN103544703A (en) * | 2013-10-19 | 2014-01-29 | 侯俊 | Digital image stitching detecting method |
| CN103914839A (en) * | 2014-03-27 | 2014-07-09 | 中山大学 | Image stitching and tampering detection method and device based on steganalysis |
| CN103955691A (en) * | 2014-05-08 | 2014-07-30 | 中南大学 | Multi-resolution LBP textural feature extracting method |
| CN103996044A (en) * | 2014-05-29 | 2014-08-20 | 天津航天中为数据系统科技有限公司 | Method and device for extracting targets through remote sensing image |
| CN103996044B (en) * | 2014-05-29 | 2017-10-10 | 天津航天中为数据系统科技有限公司 | The method and apparatus that target is extracted using remote sensing images |
| CN104244016A (en) * | 2014-08-12 | 2014-12-24 | 中山大学 | H264 video content tampering detection method |
| CN104244016B (en) * | 2014-08-12 | 2018-04-10 | 中山大学 | A kind of H264 video contents altering detecting method |
| CN104598929A (en) * | 2015-02-03 | 2015-05-06 | 南京邮电大学 | HOG (Histograms of Oriented Gradients) type quick feature extracting method |
| CN104899846A (en) * | 2015-05-20 | 2015-09-09 | 上海交通大学 | Digital image splicing passive detection method based on frequency domain local statistic model |
| CN106056523A (en) * | 2016-05-20 | 2016-10-26 | 南京航空航天大学 | Digital image stitching tampering blind detection method |
| CN106056523B (en) * | 2016-05-20 | 2019-05-24 | 南京航空航天大学 | Blind checking method is distorted in digital picture splicing |
| CN106203492A (en) * | 2016-06-30 | 2016-12-07 | 中国科学院计算技术研究所 | The system and method that a kind of image latent writing is analyzed |
| CN108645875A (en) * | 2018-03-20 | 2018-10-12 | 上海市建筑科学研究院 | A kind of defect identification method of precast shear wall grouting connection |
| CN108520215A (en) * | 2018-03-28 | 2018-09-11 | 电子科技大学 | Single sample face recognition method based on multiple dimensioned union feature encoder |
| CN108520215B (en) * | 2018-03-28 | 2022-10-11 | 电子科技大学 | Single-sample face recognition method based on multi-scale joint feature encoder |
| CN109086801A (en) * | 2018-07-06 | 2018-12-25 | 湖北工业大学 | A kind of image classification method based on improvement LBP feature extraction |
| CN109885987A (en) * | 2019-01-24 | 2019-06-14 | 中山大学 | A kind of bianry image steganalysis method based on directionality local binary patterns |
| CN109885987B (en) * | 2019-01-24 | 2023-01-24 | 中山大学 | Binary image steganography analysis method based on directional local binary pattern |
| CN111145146A (en) * | 2019-12-11 | 2020-05-12 | 北京航空航天大学 | Method and device for detecting style migration forged image based on HHT |
| CN111145146B (en) * | 2019-12-11 | 2023-04-18 | 北京航空航天大学 | Method and device for detecting style migration forged image based on HHT |
| CN111415336A (en) * | 2020-03-12 | 2020-07-14 | 泰康保险集团股份有限公司 | Image tampering identification method and device, server and storage medium |
| CN112232162A (en) * | 2020-10-06 | 2021-01-15 | 武汉烽火凯卓科技有限公司 | Pedestrian detection method and device based on multi-feature fusion cascade classifier |
| CN112232162B (en) * | 2020-10-06 | 2023-04-18 | 武汉烽火凯卓科技有限公司 | Pedestrian detection method and device based on multi-feature fusion cascade classifier |
| CN113837976A (en) * | 2021-09-17 | 2021-12-24 | 重庆邮电大学 | Multi-focus image fusion method based on combined multi-domain |
| CN113837976B (en) * | 2021-09-17 | 2024-03-19 | 重庆邮电大学 | Multi-focus image fusion method based on joint multi-domain |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103310236A (en) | Mosaic image detection method and system based on local two-dimensional characteristics | |
| He et al. | Multi-scale FCN with cascaded instance aware segmentation for arbitrary oriented word spotting in the wild | |
| Goel et al. | Dual branch convolutional neural network for copy move forgery detection | |
| Doegar et al. | Cnn based image forgery detection using pre-trained alexnet model | |
| Zhao et al. | Passive image-splicing detection by a 2-D noncausal Markov model | |
| Dev et al. | Categorization of cloud image patches using an improved texton-based approach | |
| Wang et al. | Image tampering detection based on stationary distribution of Markov chain | |
| Dua et al. | Image forgery detection based on statistical features of block DCT coefficients | |
| CN104834933A (en) | Method and device for detecting salient region of image | |
| EP3203417B1 (en) | Method for detecting texts included in an image and apparatus using the same | |
| Jaiprakash et al. | Low dimensional DCT and DWT feature based model for detection of image splicing and copy-move forgery | |
| Kalyoncu et al. | GTCLC: leaf classification method using multiple descriptors | |
| Passos et al. | A review of deep learning‐based approaches for deepfake content detection | |
| AlSawadi et al. | Copy-move image forgery detection using local binary pattern and neighborhood clustering | |
| Dua et al. | Detection and localization of forgery using statistics of DCT and Fourier components | |
| Agarwal et al. | Image forgery detection and deep learning techniques: A review | |
| CN112580480A (en) | Hyperspectral remote sensing image classification method and device | |
| Rangkuti et al. | A novel reliable approach for image batik classification that invariant with scale and rotation using MU2ECS-LBP algorithm | |
| Zagoris et al. | Scene text detection on images using cellular automata | |
| Moradi-Gharghani et al. | A new block-based copy-move forgery detection method in digital images | |
| Muhammad et al. | Image forgery detection using Gabor filters and DCT | |
| Zhao et al. | A comprehensive study on third order statistical features for image splicing detection | |
| Doegar et al. | Image forgery detection based on fusion of lightweight deep learning models | |
| CN112434731A (en) | Image recognition method and device and readable storage medium | |
| Lebedev et al. | Face detection algorithm based on a cascade of ensembles of decision trees |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20130918 |
|
| RJ01 | Rejection of invention patent application after publication |