CN107464268A - A kind of joint coding method using global and local feature - Google Patents
A kind of joint coding method using global and local feature Download PDFInfo
- Publication number
- CN107464268A CN107464268A CN201610388567.XA CN201610388567A CN107464268A CN 107464268 A CN107464268 A CN 107464268A CN 201610388567 A CN201610388567 A CN 201610388567A CN 107464268 A CN107464268 A CN 107464268A
- Authority
- CN
- China
- Prior art keywords
- mrow
- mtd
- hash
- mtr
- msub
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T9/00—Image coding
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
The present invention is a kind of joint coding method using global and local feature, including (1) the Hash perceptual coding based on DCT global characteristics；(2) the Hash coding based on SURF local features；(3) global characteristics are combined with local feature；Finally give the Hash codes of needs.The method of the present invention contrasts the single perception Hash encryption algorithm based on global characteristics and improves robustness, and contrast improves real-time using the hash algorithm of local feature merely.
Description
Technical field
It is more particularly to a kind of special using global and local the invention belongs to the network information security, image retrieval technologies field
The joint coding method of sign.
Background technology
The key technology of similar pictures search is called " perception hash algorithm " (Perceptual hash algorithm),
Its effect is to generate " fingerprint " (fingerprint) character string to every pictures, then the fingerprint of more different pictures.
As a result it is closer, just illustrate that picture is more similar.Perceptual image hash algorithm is the popular research point of rising in recent years, and traditional
Image hash algorithm is different, and it is a kind of image content-based hash algorithm to perceive hash algorithm.Its using the visual signature of image as
Basis, the regular length of image is generated by hash function or a string of fingerprints of indefinite length are used to compare.Due to perceiving Hash
Algorithm has the advantages that good real-time and adapts to big data quantity, and it is preliminary in fields such as image retrieval image authentications
Application.
Characteristics of image for perceive Hash can be divided into two kinds of global characteristics and local feature.Global characteristics include ash
Characteristic frequency characteristic color feature texture feature and line feature etc. are spent, the gray scale Hash that hash algorithm mainly has threshold value to control is calculated
Method, frequency threshold hash algorithm based on discrete cosine transform and the multidimensional global characteristics Hash combined with further feature are calculated
Method, but because global characteristics are excessively single, it is often insensitive to local features such as illumination rotational noises so that the hash algorithm resists
The not ideal enough local features of attacking ability mainly include widely used HarrisSIFT and SURF features .SIFT in recent years
Scale invariant feature conversion (Scale-invariant Feature Transform) is to be carried by David Lowe in 1999
Go out, its advantage is with yardstick illumination and rotational invariance, but because it has higher algorithm complex, in real-time system
In can not obtain applying .2006, Herbert Bay et al. to be based on image integration well and carry the methods of box filters
Acceleration version algorithm SURF (Speeded Up Robust features) the .SURF algorithms for having gone out SIFT algorithms are not subtracting significantly
In the case of few feature robustness, the speed of SIFT algorithms, which is improved, close to 3 times, turns into the preferred image local of real-time system
Perception hash algorithms of the feature extraction algorithm based on local feature is less, mainly has Hash of the Sujoy.Roy based on hyperplane to calculate
Method and V.Monga by Wavelet Feature Extraction point and carry out the hash algorithm of characteristic quantification.But global characteristics are compared, it is local
Feature still has calculating complexity, the big feature of data volume.
The content of the invention
The technical problems to be solved by the invention are the defects of overcoming prior art, there is provided one kind is special using global and local
The joint coding method of sign, the perception hash algorithm that global characteristics and local feature are combined, passes through this combination
Part real-time can be improved again by improving the characteristics of global characteristics are weak to local attack resistance.
The technical scheme is that a kind of joint coding method using global and local feature, comprises the following steps：
(1) the Hash perceptual coding based on DCT global characteristics, discrete cosine transform (DCT) are one related to Fourier
Kind conversion, for N × N images, dct transform formula is as follows：
Wherein f is N × N image slices vegetarian refreshments, and F is N × N DCT threshold matrixes, and C is cosine coefficient matrix；
Specific coding method is as follows:
A) image pixel is sampled；
B) grey scale change is carried out to coloured image, is converted into gray level image；
C) dct transform of the image is calculated, obtains transform coefficient matrix；
D) coefficient sub-matrix of the upper left corner 8 × 8 in matrix is retained；
E) average of coefficient in 8 × 8 submatrixs is calculated, is calculated as m；
F) it is as follows according to the submatrix of order traversal 8 × 8 from left to right from top to bottom, Hash rule:
Wherein x is the element in 8 × 8 submatrixs, i.e., dct transform coefficient, h (x) they are Hash codes corresponding to the element, according to
As above step can obtain the Hash coding of 64.
G) 64 2 scale codings are converted into 16 16 scale codings, the coding is the Hash perceptual coding of the image；
(2) the Hash coding based on SURF local features, it is specific as follows：
A) to keep consistent with global characteristics Hash code length, therefore image is divided into 8 × 8=64 subinterval, counted
Each subinterval SURF features points, are calculated as Fij, and i represents i-th of the subinterval of line direction from left to right herein, and j represents row side
To j-th of subinterval from top to bottom, wherein i=0,1 ..., 63；
B) traveled through according to the track of setting:
C) difference that the feature in each subinterval and last subinterval is counted is calculated, is calculated as Ei, wherein i=0,1 ..., 63；
D) differential coding principle is based on, setting Hash coding rule is as follows:
E) 64 binary system Hash codes are converted into 16 16 scale codings；
(3) global characteristics are combined with local feature, and global characteristics and local feature are combined and carry out Hash coding, it is contemplated that
The security of Hash codes transmission, randomization is carried out using following rule:
A) randomly select an integer and be designated as P, if k is 16 hexadecimal numbers, and k=P%216；
B) ring shift left that cycle-index is 16 × 4=64 is done to k and obtains 16 key k1；It is m wherein to move to left digit every time,
M is a random number in 0-9；
C) ring shift right that cycle-index is 16 × 4=64 is done to k and obtains 16 key k2；It is m wherein to move to right digit every time,
M is a random number in 0-9；
D) remember that global Hash codes are hg, local Hash codes are hl, and final Hash codes are obtained by following formula:
The beneficial effects of the present invention are：The single perception Hash coding based on global characteristics of method contrast of the present invention
Algorithm improves robustness, and contrast improves real-time using the hash algorithm of local feature merely.
Brief description of the drawings
Fig. 1 is the code track of Hash coding of the present invention based on SURF local features；
Fig. 2 is the inventive method flow chart
Embodiment
Below, carried out as described in detail below for the present invention with reference to accompanying drawing：
The method of the present invention comprises the following steps；
First, the Hash perceptual coding based on DCT global characteristics, discrete cosine transform (DCT) is one related to Fourier
Kind conversion, similar to Fourier transformation but only with real number, for N × N images, its dct transform formula is as follows：
Wherein f is N × N image slices vegetarian refreshments, and F is N × N DCT threshold matrixes, and C is cosine coefficient matrix.Connect because DCT has
The effect of nearly Karhunen-Loeve transformation, has fast algorithm again, therefore the fast algorithm for being usually used in compression of images and coding .DCT typically divides
For two kinds:Direct algorithm and Explicit Algorithm direct algorithms mainly include the matrix decomposition and recursive algorithm of dct transform；Indirectly
Algorithm mainly using DCT and DFT (discrete Fourier transform) relation, is realized, the present invention adopts by DFT fast algorithm FFT
Use the latter.The algorithm flow is as follows:
A) image pixel is sampled, is 32 × 32 pixel by image down
B) grey scale change is carried out to coloured image, is converted into gray level image
C) dct transform of the image is calculated, obtains 32 × 32 transform coefficient matrix
D) the coefficient sub-matrix of the upper left corner 8 × 8 in matrix is retained
E) average of coefficient in 8 × 8 submatrixs is calculated, is calculated as m.
F) it is as follows according to the submatrix of order traversal 8 × 8 from left to right from top to bottom, Hash rule:
Wherein x is that the element in 8 × 8 submatrixs is dct transform coefficient, and h (x) is Hash codes corresponding to the element, according to
As above step can obtain the Hash coding of 64.
G) 64 2 scale codings are converted into 16 16 scale codings, the coding is the Hash coding of the image.
Second, being encoded based on SURF local features Hash, flow is as follows：
A) in order to keep consistent with global characteristics Hash code length, therefore image is divided into 8 × 8=64 subinterval, united
Each subinterval SURF features points are counted, are calculated as Fij, i represents i-th of the subinterval of line direction from left to right herein, and j represents row
J-th of the subinterval of direction from top to bottom, wherein i=0,1 ..., 63.
B) traveled through according to Fig. 1 tracks:
C) difference that the feature in each subinterval and last subinterval is counted is calculated, is calculated as Ei, wherein i=0,1 ..., 63.
D) differential coding principle is based on, setting Hash coding rule is as follows:
E) 64 binary system Hash codes are converted into 16 16 scale codings.
The encryption algorithm has abandoned SURF description of complexity, but also remains local feature substantially and have less
Algorithm complex.
Third, the combination of global characteristics and local feature, global characteristics and local feature are now combined carry out Hash coding,
In view of the security of Hash codes transmission, randomization is carried out using following rule:
A) randomly select an integer and be designated as P, if k is 16 hexadecimal numbers, and k=P%216。
B) ring shift left that cycle-index is 16 × 4=64 is done to k and obtains 16 key k1.It is m wherein to move to left digit every time,
M is a random number in 0-9.
C) ring shift right that cycle-index is 16 × 4=64 is done to k and obtains 16 key k2.It is m wherein to move to right digit every time,
M is a random number in 0-9.
D) remember that global Hash codes are hg, local Hash codes are hl.Final Hash codes can be obtained by following formula:
Therefore finally coding H is 32 16 scale codings.
Described is only the instantiation of the present invention, any equivalent transformation based on the inventive method basis, belongs to this hair
Within bright protection domain.
Claims (1)
1. a kind of joint coding method using global and local feature, it is characterised in that comprise the following steps：
(1) the Hash perceptual coding based on DCT global characteristics, discrete cosine transform (DCT) are a kind of changes related to Fourier
Change, for N × N images, dct transform formula is as follows：
<mfenced open = "" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<mi>F</mi>
<mrow>
<mo>(</mo>
<mi>u</mi>
<mo>,</mo>
<mi>v</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mn>4</mn>
</mfrac>
<mi>C</mi>
<mrow>
<mo>(</mo>
<mi>u</mi>
<mo>)</mo>
</mrow>
<mi>C</mi>
<mrow>
<mo>(</mo>
<mi>v</mi>
<mo>)</mo>
</mrow>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>x</mi>
<mo>=</mo>
<mn>0</mn>
</mrow>
<mi>N</mi>
</munderover>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>y</mi>
<mo>=</mo>
<mn>0</mn>
</mrow>
<mi>N</mi>
</munderover>
<mi>f</mi>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>,</mo>
<mi>y</mi>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mo>(</mo>
<mn>0.1</mn>
<mo>)</mo>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mi>cos</mi>
<mrow>
<mo>(</mo>
<mfrac>
<mrow>
<mi>&pi;</mi>
<mrow>
<mo>(</mo>
<mn>2</mn>
<mi>x</mi>
<mo>+</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
<mi>u</mi>
</mrow>
<msup>
<mi>N</mi>
<mn>2</mn>
</msup>
</mfrac>
<mo>)</mo>
</mrow>
<mi>cos</mi>
<mrow>
<mo>(</mo>
<mfrac>
<mrow>
<mi>&pi;</mi>
<mrow>
<mo>(</mo>
<mn>2</mn>
<mi>y</mi>
<mo>+</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
<mi>v</mi>
</mrow>
<msup>
<mi>N</mi>
<mn>2</mn>
</msup>
</mfrac>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mo>(</mo>
<mn>0.2</mn>
<mo>)</mo>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
Wherein f is N × N image slices vegetarian refreshments, and F is N × N DCT threshold matrixes, and C is cosine coefficient matrix；
Specific coding method is as follows:
A) image pixel is sampled；
B) grey scale change is carried out to coloured image, is converted into gray level image；
C) dct transform of the image is calculated, obtains transform coefficient matrix；
D) coefficient sub-matrix of the upper left corner 8 × 8 in matrix is retained；
E) average of coefficient in 8 × 8 submatrixs is calculated, is calculated as m；
F) it is as follows according to the submatrix of order traversal 8 × 8 from left to right from top to bottom, Hash rule:
<mrow>
<mi>h</mi>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfenced open = "{" close = "">
<mtable>
<mtr>
<mtd>
<mn>1</mn>
</mtd>
<mtd>
<mrow>
<mi>i</mi>
<mi>f</mi>
</mrow>
</mtd>
<mtd>
<mrow>
<mi>x</mi>
<mo>&GreaterEqual;</mo>
<mi>m</mi>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mn>0</mn>
</mtd>
<mtd>
<mrow>
<mi>i</mi>
<mi>f</mi>
</mrow>
</mtd>
<mtd>
<mrow>
<mi>x</mi>
<mo><</mo>
<mi>m</mi>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>0.3</mn>
<mo>)</mo>
</mrow>
</mrow>
Wherein x is the element in 8 × 8 submatrixs, i.e. dct transform coefficient, and h (x) is Hash codes corresponding to the element, according to as above
Step can obtain the Hash coding of 64.
G) 64 2 scale codings are converted into 16 16 scale codings, the coding is the Hash perceptual coding of the image；
(2) the Hash coding based on SURF local features, it is specific as follows：
A) to keep consistent with global characteristics Hash code length, therefore image is divided into 8 × 8=64 subinterval, statistics is each
Subinterval SURF features points, are calculated as Fij, and i represents i-th of the subinterval of line direction from left to right herein, j represent column direction from
J-th of subinterval of top to bottm, wherein i=0,1 ..., 63；
B) traveled through according to the track of setting:
C) difference that the feature in each subinterval and last subinterval is counted is calculated, is calculated as Ei, wherein i=0,1 ..., 63；
D) differential coding principle is based on, setting Hash coding rule is as follows:
<mrow>
<mtable>
<mtr>
<mtd>
<mrow>
<mi>h</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfenced open = "{" close = "">
<mtable>
<mtr>
<mtd>
<mn>0</mn>
</mtd>
<mtd>
<mrow>
<mi>i</mi>
<mi>f</mi>
</mrow>
</mtd>
<mtd>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>0</mn>
<mo>,</mo>
<msub>
<mi>F</mi>
<mn>00</mn>
</msub>
<mo>=</mo>
<mn>0</mn>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mn>1</mn>
</mtd>
<mtd>
<mrow>
<mi>i</mi>
<mi>f</mi>
</mrow>
</mtd>
<mtd>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>0</mn>
<mo>,</mo>
<msub>
<mi>F</mi>
<mn>00</mn>
</msub>
<mo>></mo>
<mn>0</mn>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mn>0</mn>
</mtd>
<mtd>
<mrow>
<mi>i</mi>
<mi>f</mi>
</mrow>
</mtd>
<mtd>
<mrow>
<msub>
<mi>E</mi>
<mi>i</mi>
</msub>
<mo>></mo>
<mn>0</mn>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mn>1</mn>
</mtd>
<mtd>
<mrow>
<mi>i</mi>
<mi>f</mi>
</mrow>
</mtd>
<mtd>
<mrow>
<msub>
<mi>E</mi>
<mi>i</mi>
</msub>
<mo><</mo>
<mn>0</mn>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
</mrow>
</mtd>
<mtd>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>0</mn>
<mo>,</mo>
<mn>1</mn>
<mo>,</mo>
<mo>...</mo>
<mn>63</mn>
</mrow>
</mtd>
</mtr>
</mtable>
<mo>;</mo>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>0.4</mn>
<mo>)</mo>
</mrow>
</mrow>
E) 64 binary system Hash codes are converted into 16 16 scale codings；
(3) global characteristics are combined with local feature, and global characteristics and local feature are combined and carry out Hash coding, it is contemplated that Hash
The security of code transmission, randomization is carried out using following rule:
A) randomly select an integer and be designated as P, if k is 16 hexadecimal numbers, and k=P%216；
B) ring shift left that cycle-index is 16 × 4=64 is done to k and obtains 16 key k1；It is m, m 0- wherein to move to left digit every time
A random number in 9；
C) ring shift right that cycle-index is 16 × 4=64 is done to k and obtains 16 key k2；It is m, m 0- wherein to move to right digit every time
A random number in 9；
D) remember that global Hash codes are hg, local Hash codes are hl, and final Hash codes are obtained by following formula:
<mrow>
<mi>H</mi>
<mo>=</mo>
<mo>&lsqb;</mo>
<msub>
<mi>h</mi>
<mi>g</mi>
</msub>
<mo>&CirclePlus;</mo>
<msub>
<mi>k</mi>
<mn>1</mn>
</msub>
<mo>|</mo>
<msub>
<mi>h</mi>
<mn>1</mn>
</msub>
<mo>&CirclePlus;</mo>
<msub>
<mi>k</mi>
<mn>2</mn>
</msub>
<mo>&rsqb;</mo>
<mo>.</mo>
</mrow>
2
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610388567.XA CN107464268A (en) | 2016-06-02 | 2016-06-02 | A kind of joint coding method using global and local feature |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610388567.XA CN107464268A (en) | 2016-06-02 | 2016-06-02 | A kind of joint coding method using global and local feature |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107464268A true CN107464268A (en) | 2017-12-12 |
Family
ID=60545628
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610388567.XA Pending CN107464268A (en) | 2016-06-02 | 2016-06-02 | A kind of joint coding method using global and local feature |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107464268A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111078928A (en) * | 2019-12-20 | 2020-04-28 | 数据堂（北京）科技股份有限公司 | Image duplicate removal method and device |
| CN111787179A (en) * | 2020-05-30 | 2020-10-16 | 上海电力大学 | Image hash acquisition method, image security authentication method and device |
| CN113160029A (en) * | 2021-03-31 | 2021-07-23 | 海南大学 | Medical image digital watermarking method based on perceptual hashing and data enhancement |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102622366A (en) * | 2011-01-28 | 2012-08-01 | 阿里巴巴集团控股有限公司 | Similar picture identification method and similar picture identification device |
| CN103744973A (en) * | 2014-01-11 | 2014-04-23 | 西安电子科技大学 | Video copy detection method based on multi-feature Hash |
| CN105303252A (en) * | 2015-10-12 | 2016-02-03 | 国家计算机网络与信息安全管理中心 | Multi-stage nerve network model training method based on genetic algorithm |
-
2016
- 2016-06-02 CN CN201610388567.XA patent/CN107464268A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102622366A (en) * | 2011-01-28 | 2012-08-01 | 阿里巴巴集团控股有限公司 | Similar picture identification method and similar picture identification device |
| CN103744973A (en) * | 2014-01-11 | 2014-04-23 | 西安电子科技大学 | Video copy detection method based on multi-feature Hash |
| CN105303252A (en) * | 2015-10-12 | 2016-02-03 | 国家计算机网络与信息安全管理中心 | Multi-stage nerve network model training method based on genetic algorithm |
Non-Patent Citations (1)
| Title |
|---|
| 丁旭 等: "一种由DCT 和SURF 改进的图像感知哈希算法", 《小型微型计算机系统》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111078928A (en) * | 2019-12-20 | 2020-04-28 | 数据堂（北京）科技股份有限公司 | Image duplicate removal method and device |
| CN111078928B (en) * | 2019-12-20 | 2023-07-21 | 数据堂（北京）科技股份有限公司 | Image de-duplication method and device |
| CN111787179A (en) * | 2020-05-30 | 2020-10-16 | 上海电力大学 | Image hash acquisition method, image security authentication method and device |
| CN111787179B (en) * | 2020-05-30 | 2022-02-15 | 上海电力大学 | Image hash acquisition method, image security authentication method and device |
| CN113160029A (en) * | 2021-03-31 | 2021-07-23 | 海南大学 | Medical image digital watermarking method based on perceptual hashing and data enhancement |
| CN113160029B (en) * | 2021-03-31 | 2022-07-05 | 海南大学 | Medical image digital watermarking method based on perceptual hashing and data enhancement |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Liu et al. | Coverless steganography based on image retrieval of DenseNet features and DWT sequence mapping | |
| Li et al. | An Efficient Scheme for Detecting Copy-move Forged Images by Local Binary Patterns. | |
| WO2016082277A1 (en) | Video authentication method and apparatus | |
| CN110516100A (en) | A kind of calculation method of image similarity, system, storage medium and electronic equipment | |
| CN108259932B (en) | Robust hashing based on time-space domain polar coordinates cosine transform repeats video detecting method | |
| Yuan et al. | Local multi-watermarking method based on robust and adaptive feature extraction | |
| CN107464268A (en) | A kind of joint coding method using global and local feature | |
| CN104077590A (en) | Video fingerprint extraction method and system | |
| Liu et al. | Robust blind image watermarking based on chaotic mixtures | |
| Tang et al. | Robust image hashing with visual attention model and invariant moments | |
| Kumar et al. | Near lossless image compression using parallel fractal texture identification | |
| CN104361096A (en) | Image retrieval method based on characteristic enrichment area set | |
| CN101976340B (en) | License plate positioning method based on compressed domain | |
| Masotti et al. | Texture classification using invariant ranklet features | |
| CN104778689B (en) | A kind of image hashing method based on average secondary image and locality preserving projections | |
| Dhar et al. | Accurate segmentation of complex document image using digital shearlet transform with neutrosophic set as uncertainty handling tool | |
| Tang et al. | Robust image hashing via visual attention model and ring partition | |
| Zenati et al. | SSDIS-BEM: A new signature steganography document image system based on beta elliptic modeling | |
| CN104156909B (en) | The robust reversible watermark insertion of image content-based and detection method | |
| CN107358244B (en) | A kind of quick local invariant feature extracts and description method | |
| CN1246779C (en) | Embedding and extracting method for wavelet domain chaotic semi-frageile digit water print | |
| CN108090460A (en) | Expression recognition feature extraction algorithm based on multi-direction description of weber | |
| CN107609059B (en) | Chinese domain name similarity measurement method based on J-W distance | |
| CN110417551A (en) | A kind of the character picture production camouflage and restoration methods of the diffusion of combination error | |
| Walia et al. | Characterization of splicing in digital images using gray scale co-occurrence matrices |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20171212 |