CN101882310A - Method for detecting digital zoom times of digital photo - Google Patents

Method for detecting digital zoom times of digital photo Download PDF

Info

Publication number
CN101882310A
CN101882310A CN 201010196650 CN201010196650A CN101882310A CN 101882310 A CN101882310 A CN 101882310A CN 201010196650 CN201010196650 CN 201010196650 CN 201010196650 A CN201010196650 A CN 201010196650A CN 101882310 A CN101882310 A CN 101882310A
Authority
CN
China
Prior art keywords
digital
frequency
photo
zoom
zoom times
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
Application number
CN 201010196650
Other languages
Chinese (zh)
Inventor
魏为民
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shanghai University of Electric Power
University of Shanghai for Science and Technology
Original Assignee
Shanghai University of Electric Power
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Shanghai University of Electric Power filed Critical Shanghai University of Electric Power
Priority to CN 201010196650 priority Critical patent/CN101882310A/en
Publication of CN101882310A publication Critical patent/CN101882310A/en
Pending legal-status Critical Current

Links

Images

Abstract

The invention relates to a method for detecting the digital zoom times of a digital photo, which comprises the following steps of: converting the digital photo to be detected into a YCbCr space; filtering the luminance component in the digital photo by means of laplace; taking an absolute value of a filtering result to obtain a mean value according to the horizontal direction so as to obtain a vector quantity; filling zero to the tail of the vector quantity, so that the length of the vector quantity is the integral number power value of 2 to obtain a vector quantity V; converting the V by means of discrete Fourier to compute an amplitude spectrum P=log10 (1+/DFT(v)/); drawing a normalized frequency-amplitude spectrogram; and defining the digital zoom times of the photo by detecting the frequency which correspond to a peak value in the frequency-amplitude spectrogram. The method is completely not influenced by the compressing and storing quality factor of the JPEG of a digital camera, is completely independent of the EXIF information of the photo and the PRNU characteristic signal of an original digital camera, and realizes simpler and higher detection precision.

Description

The zoom times of digital detection method of digital photograph
Technical field
The present invention relates to a kind of method of zoom times of digital of photo, especially a kind of method of utilizing signal Processing and computer technology to detect the zoom times of digital of digital photograph.
Background technology
Digital zoom (Digital Zoom) is also referred to as Digital Zoom, be that eyeglass at digital camera maintains static and is in when focusing apart from state, by the processor in the digital camera, use a part of pixel on the CCD image sensor devices interpolation arithmetic to be amplified to whole image according to selected zoom multiple, as shown in Figure 1.Optical zoom (Optical Zoom) is to change the focal length of camera lens and the visual angle size of camera lens by the zoom lens set that moves forward and backward camera lens part, thus the amplification that realizes image with dwindle.Different with optical zoom, Digital Zoom is the variation that makes progress in the sensor devices vertical direction and give Zoom effect, and its real focal length does not change.The zoom times of digital of digital camera generally is no more than 6 times at present.
Though in the EXIF of digital photograph (Exchangeable Image File Format), write down the zoom times of digital raw informations such as (Digital Zoom Ratio) when taking, these information are significant for aspects such as digital evidence obtaining, multi-media information security, criminal investigation, the false news of examination, but EXIF is easy to be distorted, and makes that it is insecure only relying on EXIF information.
Existing technology detects and utilizes SVM (Support Vector Machines) and combination resampling and cfa interpolation detection can detect the convergent-divergent multiple of the non-compressed image of high-quality effectively as resampling (Resampling) detection, CFA (Color Filter Array) interpolation based on EM (Expectation Maximization) algorithm.Obtaining under the PRNU of specific digital camera (Photo-Response Non-Uniformity) the characteristic signal prerequisite possible convergent-divergent multiplying power after can detecting the photo of taking by this camera and cutting out.The whole bag of tricks at application background be not quite similar, can consult following paper:
[1]A.C.Popescu,and?H.Farid.Exposing?Digital?Forgeries?by?Detecting?Traces?ofRe-sampling[J].IEEE?Transactions?on?Signal?Processing,53(2):758-767,Feb.2005.
[2]A.C.Popescu?and?H.Farid.Exposing?Forgeries?in?Color?Filter?Array?InterpolatedImages[J].IEEE?Transactions?on?Signal?Processing,53(10):3948-3959,Oct.2005.
[3]S.Bayram,H.T.Sencar,and?N.Memon.Improvements?on?Source?Camera-modelIdentification?Based?on?CFA?Interpolation[C].Proc.of?the?WG?11.9Intl.Conference?on?Digital?Forensics,Orlando,FL,Jan.2006.
[4]A.C.Gallagher.Detection?of?Linear?and?Cubic?Interpolation?in?jpeg?compressedimages[C].IEEE?Computer?Society.Proceedings?of?the?The?2nd?CanadianConference?on?Computer?and?Robot?Vision(CRV’05),pp.65-72,Washington,DC,USA,2005.
[5]B.Mahdian?and?S.Saic.On?Periodic?Properties?of?Interpolation?and?TheirApplication?to?Image?Authentication[J].Third?Information?Symposium?onInformation?Assurance?and?Security.pp.439-446,Aug.2007.
[6]J.Buzzi?and?F.Guichard.Automatic?Detection?of?Digital?Zooms[C].TheProceedings?of?the?2004International?Conference?on?Image?Processing(ICIP2004),Singapore,Oct.2004.
[7]M.Goljan?and?J.Fridrich.Camera?Identification?from?Cropped?and?Scaled?Images[J].Security,Forensics,Steganography,and?Watermarking?of?MultimediaContents?X,Feb.2008.
Above-mentioned prior art can detect at the non-compressed image of high-quality during enforcement or only, but the digital photograph overwhelming majority is to store with the JPEG compress mode at present; Perhaps need to obtain the PRNU characteristic signal of original digital camera, realize that comparatively complexity and accuracy of detection are relatively poor.
Summary of the invention
The present invention is the zoom times of digital detection method that a kind of digital photograph will be provided, this method is not subjected to the influence of the JPEG compression storage quality factor of digital camera, do not rely on the EXIF information of photo and the PRNU characteristic signal of original digital camera fully, realize more simply, have higher accuracy of detection.
For reaching above purpose, the present invention adopts following scheme:. a kind of zoom times of digital detection method of digital photograph, concrete steps are as follows:
1. the digital photograph that desire is detected is transformed into the YCbCr space, obtains luminance component Y{i wherein, j}, i ∈ [1, M], j ∈ [1, N];
2. to luminance component Y{i, j} carries out Laplce's filtering;
3. the filtering result is taken absolute value and average, obtain vectorial U by horizontal direction;
4. the length of above-mentioned vectorial U is N, and in its end interpolation null value, the length that makes U is two integer power value, thereby obtains vectorial V;
5. V is done discrete Fourier transform (DFT), calculate amplitude spectrum: P=log 10(1+|DFT (V) |), normalized frequency-amplitude spectrogram draws;
6. by detecting the pairing frequency of peak value in frequency-amplitude spectrogram, determine the zoom times of digital of photo.
The concrete grammar of above-mentioned the 6th step is: in frequency-amplitude spectrogram, search local peaking from frequency values 1/16 to 1/2 and obtain candidate peak and respective frequencies set { P p(k), f p(k) }, rejecting set medium frequency is 1/8 ± 0.01,2/8 ± 0.01,3/8 ± 0.01 element, if set judges then that for sky the zoom times of digital of photo is 1, does not promptly pass through digital zoom; Otherwise get the wherein frequency f of peak-peak correspondence Int, calculate 1/ (1-f Int) or 1/f IntObtain the zoom times of digital of photo.
The present invention compared with prior art, have following conspicuous outstanding substantive distinguishing features and remarkable advantage: the present invention is not subjected to the influence of the JPEG compression storage quality factor of digital camera, do not rely on the EXIF information of photo and the PRNU characteristic signal of original digital camera fully, realize more simply, have higher accuracy of detection.
Description of drawings
Fig. 1 is the digital zoom synoptic diagram;
Fig. 2 is that the present invention detects block diagram;
Fig. 3 is frequency-amplitude spectrogram that embodiment produces.
Specific implementation method
Details are as follows in conjunction with the accompanying drawings for a preferred embodiment of the present invention:
As shown in Figure 2, present embodiment provides a kind of method that detects the zoom times of digital of digital photograph.To taking the digital photograph that obtains via digital zoom, after at first it being transformed into the YCbCr space, luminance component is wherein carried out Laplce's filtering, the filtering result taken absolute value obtain a vector after averaging by horizontal direction, making its length in this vector end zero padding is thereby that two integer power value obtains vectorial V, then V is done discrete Fourier transform (DFT), calculate amplitude spectrum P=log 10(1+|DFT (V) |), the normalized frequency-amplitude spectrogram that draws by detecting the pairing frequency of peak value in frequency-amplitude spectrogram, can be determined the zoom times of digital of photo.
Concrete operations step of the present invention is as follows:
By formula 1 with digital photograph to be measured (size: 2304 * 3072, camera model: Canon PowerShotA560, zoom times of digital: 1.9) be transformed into the YCbCr space, obtain luminance component Y{i wherein, j}, i ∈ [1,2304], j ∈ [1,3072].
Y Cb Cr = 16 128 128 65.481 128.553 24.966 - 37.797 - 74.203 112.000 112.000 - 93.786 - 18.214 R G B - - - ( 1 )
2. Y is carried out Laplce's filtering and obtain L{i, j}, i ∈ [1,2304], j ∈ [1,3072], wherein the Laplace operator mask is:
w = 0 1 0 1 - 4 1 0 1 0 - - - ( 2 )
To filtering as a result L take absolute value and average by horizontal direction, that is:
U ( j ) = 1 M Σ i = 1 M | L ( i , j ) | , j∈[1,3072] (3)
4. the length of above-mentioned vectorial U is 3072, and in 1024 of its end interpolations zero, the length that makes U is 2 integer power value, promptly 4096, thus obtain vectorial V.
5. (Discrete Fourier Transform DFT), calculates amplitude spectrum: P=log V to be made discrete Fourier transform (DFT) 10(1+|DFT (V) |), normalized frequency-amplitude spectrogram (left-half because symmetry is only drawn) as shown in Figure 3 draws.
6. in said frequencies-amplitude spectrogram, search local peaking from frequency values 1/16 to 1/2 and obtain candidate peak and respective frequencies set { (2.5013,0.12524), (2.7033,0.25024), (2.7942,0.37524), (2.435,0.47388) }.Here we suppose that zoom times of digital is no more than 16 times the zoom times of digital of digital camera (at present generally about 6 times), in addition because the frequency spectrum symmetry, half-and-half search and get final product.The condition of candidate peak be its value with the ratio of the average of its neighborhood medium filtering greater than threshold value T (the filter window width is 7 in can establishing generally speaking, and threshold value T is 2).After rejecting set medium frequency is the element of 1/8 ± 0.01,2/8 ± 0.01,3/8 ± 0.01 (marking with cross among Fig. 3),, promptly do not pass through digital zoom if set judges then that for sky the zoom times of digital of photo is 1; Otherwise get the wherein frequency f of peak-peak correspondence Int=0.47388 (marking with triangle among Fig. 3), by formula 4 zoom times of digital that calculate photo are 1.9007 or 2.1102:
DigitalZoomRatio = 1 1 - f int , 1 < Zoom &le; 2 1 f int , Zoom > 2 - - - ( 4 )
2.11 multiples can not occur during the zoom times of digital of considering this camera is provided with, therefore can determine that the zoom times of digital of this photo is about 1.9, consistent with the actual set value.

Claims (2)

1. the zoom times of digital detection method of a digital photograph is characterized in that concrete steps are as follows:
(1) digital photograph that desire is detected is transformed into the YCbCr space, obtains luminance component Y{i wherein, j}, i ∈ [1, M], j ∈ [1, N];
(2) to luminance component Y{i, j} carries out Laplce's filtering;
(3) the filtering result is taken absolute value average, obtain vectorial U by horizontal direction;
(4) length of above-mentioned vectorial U is N, and in its end interpolation null value, the length that makes U is two integer power value, thereby obtains vectorial V;
(5) V is done discrete Fourier transform (DFT), calculate amplitude spectrum: P=log 10(1+|DFT (V) |), normalized frequency-amplitude spectrogram draws;
(6) by detecting the pairing frequency of peak value in frequency-amplitude spectrogram, determine the zoom times of digital of photo.
2. the zoom times of digital detection method of digital photograph according to claim 1, it is characterized in that: the concrete grammar of above-mentioned the 6th step is: in frequency-amplitude spectrogram, search local peaking from frequency values 1/16 to 1/2 and obtain candidate peak and respective frequencies set { P p(k), f p(k) }, rejecting set medium frequency is 1/8 ± 0.01,2/8 ± 0.01,3/8 ± 0.01 element, if set judges then that for sky the zoom times of digital of photo is 1, does not promptly pass through digital zoom; Otherwise get the wherein frequency f of peak-peak correspondence Int, calculate 1/ (1-f Int) or l/f IntObtain the zoom times of digital of photo.
CN 201010196650 2010-06-10 2010-06-10 Method for detecting digital zoom times of digital photo Pending CN101882310A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN 201010196650 CN101882310A (en) 2010-06-10 2010-06-10 Method for detecting digital zoom times of digital photo

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN 201010196650 CN101882310A (en) 2010-06-10 2010-06-10 Method for detecting digital zoom times of digital photo

Publications (1)

Publication Number Publication Date
CN101882310A true CN101882310A (en) 2010-11-10

Family

ID=43054321

Family Applications (1)

Application Number Title Priority Date Filing Date
CN 201010196650 Pending CN101882310A (en) 2010-06-10 2010-06-10 Method for detecting digital zoom times of digital photo

Country Status (1)

Country Link
CN (1) CN101882310A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105865327A (en) * 2015-01-22 2016-08-17 成都飞机工业(集团)有限责任公司 Zoom photograph-based relative position measurement method
CN116735167A (en) * 2023-08-16 2023-09-12 深圳市灿锐科技有限公司 Lens multiplying power and distortion detection method

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1527588A (en) * 2003-03-04 2004-09-08 ������������ʽ���� Signal processing device and image data generating device
CN1799068A (en) * 2003-07-08 2006-07-05 佳能株式会社 Image registration method improvement
JP2006195023A (en) * 2005-01-12 2006-07-27 Sony Corp Imaging apparatus
EP1713017A1 (en) * 2005-04-11 2006-10-18 Valeo Vision Procedure, device and camera for detection of objects using digital images
CN101477678A (en) * 2008-12-04 2009-07-08 上海大学 Blind detection method for digital image scaling factor

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1527588A (en) * 2003-03-04 2004-09-08 ������������ʽ���� Signal processing device and image data generating device
CN1799068A (en) * 2003-07-08 2006-07-05 佳能株式会社 Image registration method improvement
JP2006195023A (en) * 2005-01-12 2006-07-27 Sony Corp Imaging apparatus
EP1713017A1 (en) * 2005-04-11 2006-10-18 Valeo Vision Procedure, device and camera for detection of objects using digital images
CN101477678A (en) * 2008-12-04 2009-07-08 上海大学 Blind detection method for digital image scaling factor

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105865327A (en) * 2015-01-22 2016-08-17 成都飞机工业(集团)有限责任公司 Zoom photograph-based relative position measurement method
CN116735167A (en) * 2023-08-16 2023-09-12 深圳市灿锐科技有限公司 Lens multiplying power and distortion detection method
CN116735167B (en) * 2023-08-16 2023-10-13 深圳市灿锐科技有限公司 Lens multiplying power and distortion detection method

Similar Documents

Publication Publication Date Title
Feng et al. Normalized energy density-based forensic detection of resampled images
US10540806B2 (en) Systems and methods for depth-assisted perspective distortion correction
Yerushalmy et al. Digital image forgery detection based on lens and sensor aberration
US8249382B2 (en) Image-capturing apparatus, image processing method and program product
Birajdar et al. Digital image forgery detection using passive techniques: A survey
Luo et al. A survey of passive technology for digital image forensics
Ji et al. Robust video restoration by joint sparse and low rank matrix approximation
US8055900B2 (en) Information processing apparatus, information processing system, information processing method, storage medium and program
EP2323374B1 (en) Image pickup apparatus, image pickup method, and program
JP4948593B2 (en) Composite detection in previously compressed images
JPWO2005093653A1 (en) Image correction apparatus and method, image correction database creating method, information data providing apparatus, image processing apparatus, information terminal, and information database apparatus
WO2012075817A1 (en) Text image trimming method
CN104392416A (en) Video stitching method for sports scene
CN102957915B (en) Double JPEG (Joint Photographic Experts Group) compressed image-targeted tamper detection and tamper locating method
Wahab et al. Inter-camera model image source identification with conditional probability features
Abdulrahman et al. Color image stegananalysis using correlations between RGB channels
US20130208984A1 (en) Content scene determination device
CN101477678B (en) Blind detection method for digital image scaling factor
CN101882310A (en) Method for detecting digital zoom times of digital photo
CN109729231A (en) A kind of file scanning method, device and equipment
WO2009065151A2 (en) Method for image quality assessment using quality vectors
Rani et al. Digital image forgery detection techniques: A comprehensive review
JP2004185386A (en) Image collating unit
Gharibi et al. Using the local information of image to identify the source camera
Lukac Single-sensor imaging in consumer digital cameras: a survey of recent advances and future directions

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C02 Deemed withdrawal of patent application after publication (patent law 2001)
WD01 Invention patent application deemed withdrawn after publication

Application publication date: 20101110