CN111428740A - Detection method and device for network-shot photo, computer equipment and storage medium - Google Patents
Detection method and device for network-shot photo, computer equipment and storage medium Download PDFInfo
- Publication number
- CN111428740A CN111428740A CN202010127392.3A CN202010127392A CN111428740A CN 111428740 A CN111428740 A CN 111428740A CN 202010127392 A CN202010127392 A CN 202010127392A CN 111428740 A CN111428740 A CN 111428740A
- Authority
- CN
- China
- Prior art keywords
- fourier
- picture
- video
- identified
- peak value
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 59
- 230000009466 transformation Effects 0.000 claims abstract description 117
- 238000000034 method Methods 0.000 claims description 40
- 238000002372 labelling Methods 0.000 claims description 19
- 238000004590 computer program Methods 0.000 claims description 16
- 238000012549 training Methods 0.000 claims description 14
- 238000000605 extraction Methods 0.000 claims description 6
- 230000006870 function Effects 0.000 description 7
- 238000010586 diagram Methods 0.000 description 6
- 238000001228 spectrum Methods 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 2
- 238000013145 classification model Methods 0.000 description 2
- 239000002131 composite material Substances 0.000 description 2
- 238000013403 standard screening design Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/42—Global feature extraction by analysis of the whole pattern, e.g. using frequency domain transformations or autocorrelation
- G06V10/431—Frequency domain transformation; Autocorrelation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computing Systems (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a detection method and device for a network-shot photo, computer equipment and a storage medium, which are applied to the technical field of picture identification and used for solving the technical problem of high error identification rate of identifying whether a video or a photo is a shot photo in the prior art. The detection method of the network copied photo provided by the invention comprises the following steps: when a picture or a video to be identified is received, extracting a Fourier characteristic value of the picture or extracting a Fourier characteristic value of a video frame picture of the video; generating a Fourier feature map of the picture or the video frame picture according to the extracted Fourier feature values; identifying the number of peak value transformation points in the Fourier characteristic map through a pre-trained target detection model; and when the number of the identified peak value transformation points is at least two and the at least two peak value transformation points are symmetrically distributed, judging that the picture or the video to be identified corresponding to the Fourier feature map is a reproduction.
Description
Technical Field
The invention relates to the technical field of picture recognition, in particular to a method and a device for detecting a network-shot photo, computer equipment and a storage medium.
Background
Based on the needs of some scenes, it is necessary to identify whether a picture or video is a copied picture or video. Whether people's eye discernment certain picture or video exist great error for the reproduction, and the prior art of present intelligent recognition picture or video whether for the reproduction mainly includes following two kinds of schemes:
(1) the method comprises the steps of generating a picture or a video frame through a picture copied on a screen, then segmenting the generated picture or video frame, extracting characteristic value identification pixel points and moire fringes, and judging whether the corresponding picture or video is copied or not according to the characteristic value identification pixel points and the moire fringes. However, the disadvantage of this scheme is that an original copied image is needed, the image size is large, so that the data computation amount is large, the processing speed is slow, and the shooting angle and distance have certain requirements, so that if the image is not shot as required, pixels with a large probability and moire fringes are not obvious, which results in that the identification cannot be achieved, the error identification rate is high, and if the shot object with a large number of similar pixels is identified as the copied image, the error identification rate is improved.
(2) Extracting Fourier spectrum characteristics of RGB three channels of a file to be identified, inputting the extracted characteristics into a classification model, and determining whether the picture or video to be identified is a copy through the classification model. The existing identification method needs to extract the Fourier spectrum characteristics of RGB three channels of a picture or a video frame to be identified, so that the calculation amount is large, and if a shot object with a large number of similar pixel points is shot, the scheme can identify the picture or the video to be identified as a copy, so that the identification result is inaccurate.
Disclosure of Invention
The embodiment of the invention provides a method and a device for detecting a network-shot photo, a computer device and a storage medium, which aim to solve the technical problem of high error identification rate of identifying whether a video or a photo is a copy in the prior art.
A method for detecting a network-shot photo, the method comprising:
when a picture or a video to be identified is received, extracting a Fourier characteristic value of the picture or extracting a Fourier characteristic value of a video frame picture of the video;
generating a Fourier feature map of the picture or the video frame picture according to the extracted Fourier feature values;
identifying the number of peak value transformation points in the Fourier characteristic map through a pre-trained target detection model;
and when the number of the identified peak value transformation points is at least two and the at least two peak value transformation points are symmetrically distributed, judging that the picture or the video to be identified corresponding to the Fourier feature map is a reproduction.
A device for detecting network-shot photographs, the device comprising:
the extraction module is used for extracting a Fourier characteristic value of the picture or extracting a Fourier characteristic value of a video frame picture of the video when the picture or the video to be identified is received;
the generating module is used for generating a Fourier feature map of the picture or the video frame picture according to the extracted Fourier feature value;
the identification module is used for identifying the number of peak value transformation points in the Fourier characteristic map through a pre-trained target detection model;
and the judging module is used for judging that the picture or video to be identified corresponding to the Fourier feature map is a reproduction when the number of the identified peak value transformation points is at least two and the at least two peak value transformation points are symmetrically distributed.
A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above detection method for network-shot photos when executing the computer program.
A computer-readable storage medium, which stores a computer program, which, when executed by a processor, implements the steps of the above-described method for detecting network-captured photographs.
According to the detection method, device, computer equipment and storage medium for the network-shot photos, the Fourier characteristic value of the picture is extracted or the Fourier characteristic value of the video frame picture of the video is extracted, then the picture or the Fourier characteristic map of the video frame picture is generated according to the extracted Fourier characteristic value, the number of peak value transformation points in the Fourier characteristic map is identified through a pre-trained target detection model, and finally whether the corresponding picture or photo is shot is judged according to the identified number of the peak value transformation points And shooting is carried out, so that the accuracy of the copying identification is improved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the description of the embodiments of the present invention will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to these drawings without inventive labor.
FIG. 1 is a schematic diagram of an application environment of a method for detecting a network-captured photo according to an embodiment of the present invention;
FIG. 2 is a flowchart of a method for detecting network-captured photographs according to an embodiment of the present invention;
FIG. 3 is a flowchart of a method for detecting network-captured photos according to another embodiment of the present invention;
FIG. 4 is a flowchart of another method for detecting network-captured photographs according to an embodiment of the present invention;
FIG. 5-a illustrates the location of peak transform points in a non-replicated picture or video in accordance with an embodiment of the present invention;
FIG. 5-b is a diagram of the location of peak transform points in a rendered picture or video in accordance with an embodiment of the present invention;
FIG. 6 is a schematic structural diagram of an apparatus for detecting network-captured photos according to an embodiment of the present invention;
FIG. 7 is a schematic diagram of a computer device according to an embodiment of the invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The detection method for the network-shot photo can be applied to the application environment shown in fig. 1. The computer devices shown in fig. 1 include, but are not limited to, various personal computers, notebook computers, smart phones, tablet computers, and portable wearable devices.
Fig. 2 is a flowchart of a method for detecting a network-captured photo according to an embodiment of the present invention, and in an embodiment, as shown in fig. 2, a method for detecting a network-captured photo is provided, which is described by taking the computer device in fig. 1 as an example, and includes the following steps S101 to S104.
S101, when a picture or a video to be identified is received, extracting a Fourier characteristic value of the picture or extracting a Fourier characteristic value of a video frame picture of the video.
In one embodiment, when the video to be identified is received, the step further comprises:
randomly extracting a plurality of video frame pictures from the video to be identified;
and extracting the Fourier characteristic value of each video frame picture.
In one embodiment, the received picture or video to be identified includes an original picture or video and also includes a compressed picture or video.
In this embodiment, the video frame picture may be any one frame or any multiple frames in the video to be identified.
S102, generating a Fourier feature map of the picture or the video frame picture according to the extracted Fourier feature value.
In one embodiment, when the video to be identified is received, the step further comprises:
and generating a Fourier feature map of each video frame picture in the same video to be identified according to the extracted Fourier feature values.
S103, identifying the number of peak value transformation points in the Fourier feature map through a pre-trained target detection model.
In one embodiment, the target detection model may be an SSD (Single Shot multi box detector) target detection model, which is an end-to-end Single-stage target detection network that treats the detected target on the picture as a discretized composite of default boxes of a number of default boxes.
Generally, the picture to be detected includes targets with different sizes and resolutions. To effectively cover such target detection at a variety of different scales and resolutions, SSDs expand the range of feature choices.
And S104, when the number of the identified peak value transformation points is at least two and the at least two peak value transformation points are symmetrically distributed, judging that the picture or video to be identified corresponding to the Fourier feature map is a reproduction.
In one embodiment, when the video to be identified is received, the step further comprises:
and when the number of the peak value transformation points of the corresponding Fourier feature map in each video frame picture included in the same video to be identified is at least two, judging that the corresponding video to be identified is a reproduction.
Fig. 5-a is a position of a peak transformation point in a non-copied picture or video in an embodiment of the present invention, and fig. 5-b is a position of a peak transformation point in a copied picture or video in an embodiment of the present invention, as shown by comparing fig. 5-a with fig. 5-b, it is found through multiple tests that the number of peak transformation points in the non-copied picture or video is 1, and the number of peak transformation points in the copied picture or video is generally many, so that the number of peak transformation points in the fourier feature map can be identified by extracting the fourier feature map of the picture or video to be identified, and by using a pre-trained target detection model, and whether the corresponding picture or video is copied is determined according to the number of peak transformation points.
In the embodiment, the Fourier characteristic value of the picture is extracted or the Fourier characteristic value of the video frame picture of the video is extracted, then the Fourier characteristic map of the picture or the video frame picture is generated according to the extracted Fourier characteristic value, the number of peak value transformation points in the Fourier characteristic map is identified through a pre-trained target detection model, and finally whether the corresponding picture or photo is a reproduction is judged according to the identified number of the peak value transformation points, compared with the existing reproduction identification technology, the method does not need to extract the Fourier spectrum characteristics of three channels of RGB (red, green and blue), but directly generates the Fourier characteristic map, reduces the calculation amount of reproduction identification, improves the operation speed of reproduction identification, and on the other hand, judges whether the corresponding picture or photo is a reproduction according to the identified number of the peak value transformation points, the accuracy of the copying identification is improved.
Fig. 3 is a flowchart of a method for detecting network-captured photos according to another embodiment of the present invention, and the step of training the object detection model according to the present embodiment is shown in fig. 3, wherein the step of training the object detection model includes the following steps S301 to S304.
S301, extracting the Fourier characteristic value of the copied sample picture, and generating a copied Fourier characteristic map according to the Fourier characteristic value of the copied sample picture.
In one embodiment, the extracted fourier feature value of the copied sample picture is mainly the texture feature of the sample picture, and the texture feature of the copied picture is more prominent than the texture feature of a non-copied picture, and is the texture of a picture on a display screen of a mobile phone, a computer and the like which is directly copied.
The copied sample picture comprises a picture of a copied finished photo, a picture displayed on a copied screen interface, a picture of a video picture played in a copied screen and the like.
S302, extracting the Fourier characteristic value of the non-copied sample picture, and generating a non-copied Fourier characteristic map according to the Fourier characteristic value of the non-copied sample picture.
In this embodiment, the Fourier signature is as shown in the right hand picture of FIG. 5-a or 5-b.
S303, receiving a first labeling area of a peak value transformation point in the copied Fourier feature map, and receiving a second labeling area of a peak value transformation point in the non-copied Fourier feature map.
In one embodiment, the first labeled region and the second labeled region are regions input by a user for machine learning which shape belongs to a peak transformation point.
S304, inputting the copied Fourier feature map, the first labeling area, the non-copied Fourier feature map and the second labeling area into the target detection model for training to obtain the trained target detection model.
The training process is a process of enabling the machine to learn what shape belongs to the 'peak value transformation point', the machine adjusts parameters of the target detection model through continuously learning the area of the peak value transformation point marked by the user in the training process to obtain the trained target detection model, and the peak value transformation point in the unknown Fourier characteristic map can be automatically detected.
In one embodiment, the target detection model may be an SSD (Single Shot multiple boxdetector, deep learning based target detection algorithm) model.
According to a training scenario of this embodiment, for example, a fourier feature map image with a first labeled region/a second labeled region may be input to a third convolutional layer (conv5_3) in a fifth convolutional block in an vgg _16 neural network of a target detection model SSD, a feature map feature _ map (h _ w _ c) is obtained, the feature map is taken as a plurality of windows, a receptive field of the window is calculated, a corresponding default box is generated through the calculated receptive field, all the first labeled regions and the second labeled regions are traversed, a position offset of convolutional sampling is learned and adjusted, so that a predicted peak transform point region prediction box includes an actual peak transform point labeling region default box, until a maximum prediction range iou (iterative unit) is obtained through output, and training is completed.
The embodiment provides a method for training the target detection model, the target detection model which automatically identifies the peak value transformation point is trained in advance to identify the peak value transformation point in the Fourier map, so that the identification speed of the picture or video to be identified can be increased, and the identification efficiency is higher.
Fig. 4 is a flowchart of another method for detecting a network-captured photo according to an embodiment of the present invention, and the method for detecting a network-captured photo according to an embodiment of the present invention is described in detail below with reference to fig. 4, and as shown in fig. 4, the method further includes the following steps S401 to S403 based on the steps S101 and S102.
S401, identifying the number and the positions of peak value transformation points in the Fourier feature map through a pre-trained target detection model.
S402, when the number of the identified peak value transformation points is at least two, identifying a central peak value transformation point and a non-central peak value transformation point according to the positions of the peak value transformation points.
In one embodiment, the peak transform point located at the center position in the fourier feature map may be determined as the center peak transform point, and all other peak transform points may be determined as non-peak transform points.
And S403, judging whether the non-peak value transformation points are symmetrically distributed around the central peak value transformation point, if so, judging that the picture or video to be identified corresponding to the Fourier feature map is a copy, and otherwise, judging that the corresponding picture or video to be identified is a non-copy.
In one embodiment, as shown in fig. 5-a by comparing with fig. 5-b, further analyzing the positions of the peak transformation points in the copied picture or video shown in fig. 5-b, it can be known that the non-central peak transformation points are distributed in central symmetry around the central peak transformation point, and therefore, this feature can be used as a further limiting condition for determining whether the corresponding picture or video to be identified is copied, so as to further improve the accuracy of the copy identification.
In one embodiment, when a video to be identified is received, the method for detecting a network-shot photo further includes:
identifying the position of a peak value transformation point in each Fourier feature map in the same video to be identified through the target detection model;
and when the number of the peak value transformation points in each Fourier feature map in the same identified video to be identified is at least two, identifying the central peak value transformation point and the non-central peak value transformation point of each Fourier feature map according to the positions of the peak value transformation points.
In this embodiment, the step of determining that the to-be-identified picture or video corresponding to the fourier feature map is a copy in the above step further includes:
and judging whether the non-peak value transformation points in each Fourier feature map are symmetrically distributed around the corresponding central peak value transformation point, if so, judging that the corresponding video to be identified is a copy, and otherwise, judging that the corresponding video to be identified is a non-copy.
In this embodiment, when an object to be identified is defined as a video, a plurality of video frame pictures are arbitrarily extracted from the video, position identification of peak transformation points in a fourier feature map, judgment of the number of peak transformation points, and judgment of whether non-central peak transformation points are symmetrically distributed around the central peak transformation point are performed on each frame picture belonging to the same video to be identified, and when the non-peak transformation points in each fourier feature map are symmetrically distributed around the corresponding central peak transformation point, it is judged that the corresponding video to be identified is a reprint, so as to improve identification accuracy of the reprint video.
It should be understood that, the sequence numbers of the steps in the foregoing embodiments do not imply an execution sequence, and the execution sequence of each process should be determined by its function and inherent logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.
Fig. 6 is a schematic structural diagram of a device for detecting a network-captured photo according to an embodiment of the present invention, and the device for detecting a network-captured photo according to an embodiment of the present invention is described in detail below with reference to fig. 6, where the device for detecting a network-captured photo provided in fig. 6 corresponds to the method for detecting a network-captured photo according to the embodiment. As shown in fig. 6, the apparatus 100 for detecting network-shot photos includes an extraction module 11, a generation module 12, an identification module 13 and a judgment module 14.
The extraction module 11 is configured to extract a fourier feature value of the picture or extract a fourier feature value of a video frame picture of the video when the picture or the video to be identified is received.
In one embodiment, the received picture or video to be identified includes an original picture or video and also includes a compressed picture or video.
In this embodiment, the video frame picture may be any one frame or any multiple frames in the video to be identified.
A generating module 12, configured to generate a fourier feature map of the picture or the video frame picture according to the extracted fourier feature value.
And the identification module 13 is configured to identify the number of peak value transformation points in the fourier feature map through a pre-trained target detection model.
In one embodiment, the target detection model may be an SSD (Single Shot multi box detector) target detection model, which is an end-to-end Single-stage target detection network that treats the detected target on the picture as a discretized composite of default boxes of a number of default boxes.
Generally, the picture to be detected includes targets with different sizes and resolutions. To effectively cover such target detection at a variety of different scales and resolutions, SSDs expand the range of feature choices.
And the judging module 14 is configured to judge that the picture or video to be identified corresponding to the fourier feature map is a copy when the number of the identified peak transformation points is at least two and the at least two peak transformation points are symmetrically distributed.
Fig. 5-a is a position of a peak transformation point in a non-copied picture or video in an embodiment of the present invention, and fig. 5-b is a position of a peak transformation point in a copied picture or video in an embodiment of the present invention, as shown by comparing fig. 5-a with fig. 5-b, it is found through multiple tests that the number of peak transformation points in the non-copied picture or video is 1, and the number of peak transformation points in the copied picture or video is generally many, so that the fourier feature map of the picture or video to be identified can be extracted by the extraction module 11, the identification module 13 identifies the number of peak transformation points in the fourier feature map through a pre-trained target detection model, and finally, the judgment module 14 judges whether the corresponding picture or video is copied according to the number of peak transformation points.
In one embodiment, the apparatus 100 for detecting network-shot photos further includes:
the first generation unit is used for extracting the Fourier characteristic value of the copied sample picture and generating a copied Fourier characteristic map according to the Fourier characteristic value of the copied sample picture;
the second generation unit is used for extracting the Fourier characteristic value of the non-copied sample picture and generating a non-copied Fourier characteristic map according to the Fourier characteristic value of the non-copied sample picture;
the labeling module is used for receiving a first labeling area of a peak value transformation point in the copied Fourier feature map and receiving a second labeling area of the peak value transformation point in the non-copied Fourier feature map;
and the training module is used for inputting the copied Fourier feature map, the first labeling area, the non-copied Fourier feature map and the second labeling area into the target detection model for training to obtain the trained target detection model.
Further, in one embodiment, the apparatus for detecting a network-shot photo further includes:
the position identification module is used for identifying the position of a peak value transformation point in the Fourier feature map through the target detection model; the central peak value transformation point and the non-central peak value transformation point are identified according to the positions of the peak value transformation points when the number of the identified peak value transformation points is at least two;
the judgment module is specifically configured to judge whether the non-peak transformation points are symmetrically distributed around the central peak transformation point, if so, judge that the picture or video to be identified corresponding to the fourier feature map is a copy, otherwise, judge that the corresponding picture or video to be identified is a non-copy.
In this embodiment, as shown in fig. 5-a by comparing with fig. 5-b, further analyzing the positions of the peak transformation points in the copied picture or video shown in fig. 5-b, it can be known that the non-central peak transformation points are distributed in central symmetry around the central peak transformation point, and therefore, this feature can be used as a further limiting condition for determining whether the corresponding picture or video to be identified is copied, so as to further improve the accuracy of the copied identification.
In one embodiment, when a video to be identified is received:
the extraction module is specifically configured to extract a plurality of video frame pictures from the video to be identified at will, and is further configured to extract a fourier feature value of each of the video frame pictures.
The generating module is specifically configured to generate a fourier feature map of each video frame picture in the same video to be identified according to the extracted fourier feature values.
The identification module is specifically used for identifying the number of peak value transformation points in each Fourier feature map in the same video to be identified.
The judgment module is specifically used for judging that the corresponding video to be identified is a reproduction when the number of peak value transformation points of the corresponding Fourier feature map in each video frame picture included in the same video to be identified is at least two.
In this embodiment, when an object to be identified is defined as a video, a plurality of video frame pictures are arbitrarily extracted from the video, position identification of peak transformation points in a fourier feature map, judgment of the number of peak transformation points, and judgment of whether non-central peak transformation points are symmetrically distributed around the central peak transformation point are performed on each frame picture belonging to the same video to be identified, and when the non-peak transformation points in each fourier feature map are symmetrically distributed around the corresponding central peak transformation point, it is judged that the corresponding video to be identified is a reprint, so as to improve identification accuracy of the reprint video.
For specific limitations of the detection device for network-captured photos, reference may be made to the above limitations on the detection method for network-captured photos, which are not described herein again. All or part of the modules in the detection device for network-shot photos can be realized by software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as shown in fig. 7. The computer device includes a processor, a memory, a network interface, a display screen, and an input device connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The network interface of the computer device is used for communicating with an external server through a network connection. The computer program is executed by a processor to implement a method for detecting network-shot photos.
In one embodiment, a computer device is provided, which includes a memory, a processor, and a computer program stored on the memory and executable on the processor, and when the processor executes the computer program, the steps of the detection method for network-shot photos in the above embodiments are implemented, for example, steps 101 to 104 shown in fig. 2. Alternatively, the processor, when executing the computer program, implements the functions of the modules/units of the detection apparatus for network-captured photos in the above-described embodiments, such as the functions of the modules 11 to 14 shown in fig. 7. To avoid repetition, further description is omitted here.
In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and the computer program is executed by a processor to implement the steps of the detection method for network-shot photos in the above-described embodiments, such as the steps 101 to 104 shown in fig. 2. Alternatively, the computer program, when executed by the processor, implements the functions of the modules/units of the detection apparatus for network-captured photos in the above-described embodiments, such as the functions of the modules 11 to 14 shown in fig. 7. To avoid repetition, further description is omitted here.
It will be understood by those of ordinary skill in the art that all or a portion of the processes of the methods of the embodiments described above may be implemented by a computer program that may be stored on a non-volatile computer-readable storage medium, which when executed, may include the processes of the embodiments of the methods described above, wherein any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory.
In the method, the apparatus, the computer device, and the storage medium for detecting a network-captured photo provided by this embodiment, the fourier feature value of the picture is extracted or the fourier feature value of the video frame picture of the video is extracted, then the fourier feature map of the picture or the video frame picture is generated according to the extracted fourier feature value, then the number of peak transform points in the fourier feature map is identified by a pre-trained target detection model, and finally whether the corresponding picture or photo is captured in a captured state is determined according to the identified number of the peak transform points, compared with the existing capturing identification technology, in this application, the fourier feature map is directly generated without extracting the fourier spectrum features of three RGB (red, green, and blue) channels, so that the amount of computation of capturing identification is reduced, and the operation speed of capturing identification is improved, on the other hand, whether the corresponding picture or photo is a reproduction or not is judged according to the identified number of the peak value transformation points, and the reproduction identification accuracy is improved.
It will be apparent to those skilled in the art that, for convenience and brevity of description, only the above-mentioned division of the functional units and modules is illustrated, and in practical applications, the above-mentioned function distribution may be performed by different functional units and modules according to needs, that is, the internal structure of the apparatus is divided into different functional units or modules to perform all or part of the above-mentioned functions.
The above-mentioned embodiments are only used for illustrating the technical solutions of the present invention, and not for limiting the same; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; such modifications and substitutions do not substantially depart from the spirit and scope of the embodiments of the present invention, and are intended to be included within the scope of the present invention.
Claims (10)
1. A detection method for network-shot photos is characterized by comprising the following steps:
when a picture or a video to be identified is received, extracting a Fourier characteristic value of the picture or extracting a Fourier characteristic value of a video frame picture of the video;
generating a Fourier feature map of the picture or the video frame picture according to the extracted Fourier feature values;
identifying the number of peak value transformation points in the Fourier characteristic map through a pre-trained target detection model;
and when the number of the identified peak value transformation points is at least two and the at least two peak value transformation points are symmetrically distributed, judging that the picture or the video to be identified corresponding to the Fourier feature map is a reproduction.
2. The method for detecting the network-shot photo according to claim 1, wherein the step of training the target detection model comprises:
extracting a Fourier characteristic value of the copied sample picture, and generating a copied Fourier characteristic map according to the Fourier characteristic value of the copied sample picture;
extracting a Fourier characteristic value of the non-copied sample picture, and generating a non-copied Fourier characteristic map according to the Fourier characteristic value of the non-copied sample picture;
receiving a first labeling area of a peak value transformation point in the copied Fourier feature map, and receiving a second labeling area of a peak value transformation point in the non-copied Fourier feature map;
inputting the copied Fourier feature map, the first labeling area, the non-copied Fourier feature map and the second labeling area into the target detection model for training to obtain the trained target detection model.
3. The method for detecting the network-shot photo according to claim 1, wherein the method further comprises:
identifying, by the target detection model, locations of peak transform points in the Fourier feature map;
the step of judging that the picture or video to be identified corresponding to the Fourier feature map is a reproduction further comprises:
when the number of the identified peak value transformation points is at least two, identifying a central peak value transformation point and a non-central peak value transformation point according to the positions of the peak value transformation points;
and judging whether the non-peak value transformation points are symmetrically distributed around the central peak value transformation point, if so, judging that the picture or video to be identified corresponding to the Fourier characteristic map is a copy, and otherwise, judging that the corresponding picture or video to be identified is a non-copy.
4. The method for detecting the network-shot photo according to claim 1, wherein when the video to be identified is received, the step of extracting the fourier feature value of the picture or extracting the fourier feature value of the video frame picture of the video comprises:
randomly extracting a plurality of video frame pictures from the video to be identified;
extracting a Fourier characteristic value of each video frame picture;
the step of generating a fourier feature map of the picture or the video frame picture according to the extracted fourier feature values further comprises:
generating a Fourier feature map of each video frame picture in the same video to be identified according to the extracted Fourier feature values;
when the number of the identified peak value transformation points is at least two, the step of judging that the picture or the video to be identified corresponding to the Fourier feature map is a reproduction further comprises the following steps:
and when the number of the peak value transformation points of the corresponding Fourier feature map in each video frame picture included in the same video to be identified is at least two, judging that the corresponding video to be identified is a reproduction.
5. The method for detecting the network-shot photo according to claim 4, wherein the method further comprises:
identifying the position of a peak value transformation point in each Fourier feature map in the same video to be identified through the target detection model;
when the number of the peak value transformation points in each Fourier feature map in the same identified video to be identified is at least two, identifying a central peak value transformation point and a non-central peak value transformation point of each Fourier feature map according to the positions of the peak value transformation points;
the step of judging that the picture or video to be identified corresponding to the Fourier feature map is a reproduction further comprises:
and judging whether the non-peak value transformation points in each Fourier feature map are symmetrically distributed around the corresponding central peak value transformation point, if so, judging that the corresponding video to be identified is a copy, and otherwise, judging that the corresponding picture or video to be identified is a non-copy.
6. A device for detecting network-shot photos, said device comprising:
the extraction module is used for extracting a Fourier characteristic value of the picture or extracting a Fourier characteristic value of a video frame picture of the video when the picture or the video to be identified is received;
the generating module is used for generating a Fourier feature map of the picture or the video frame picture according to the extracted Fourier feature value;
the identification module is used for identifying the number of peak value transformation points in the Fourier characteristic map through a pre-trained target detection model;
and the judging module is used for judging that the picture or video to be identified corresponding to the Fourier feature map is a reproduction when the number of the identified peak value transformation points is at least two and the at least two peak value transformation points are symmetrically distributed.
7. The apparatus for detecting network-shot photos according to claim 6, wherein said apparatus for detecting network-shot photos further comprises:
the first generation unit is used for extracting the Fourier characteristic value of the copied sample picture and generating a copied Fourier characteristic map according to the Fourier characteristic value of the copied sample picture;
the second generation unit is used for extracting the Fourier characteristic value of the non-copied sample picture and generating a non-copied Fourier characteristic map according to the Fourier characteristic value of the non-copied sample picture;
the labeling module is used for receiving a first labeling area of a peak value transformation point in the copied Fourier feature map and receiving a second labeling area of the peak value transformation point in the non-copied Fourier feature map;
and the training module is used for inputting the copied Fourier feature map, the first labeling area, the non-copied Fourier feature map and the second labeling area into the target detection model for training to obtain the trained target detection model.
8. The apparatus for detecting network-shot photos according to claim 6, wherein said apparatus for detecting network-shot photos further comprises:
the position identification module is used for identifying the positions of peak value transformation points in the Fourier characteristic map through the target detection model, and is also used for identifying a central peak value transformation point and a non-central peak value transformation point according to the positions of the peak value transformation points when the number of the identified peak value transformation points is at least two;
the judgment module is specifically configured to judge whether the non-peak transformation points are symmetrically distributed around the central peak transformation point, if so, judge that the picture or video to be identified corresponding to the fourier feature map is a copy, otherwise, judge that the corresponding picture or video to be identified is a non-copy.
9. Computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that the processor implements the steps of the method for detecting network-captured photographs according to any one of claims 1 to 5 when executing the computer program.
10. A computer-readable storage medium, in which a computer program is stored, which, when being executed by a processor, carries out the steps of the method for detecting network-captured photographs according to any one of claims 1 to 5.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010127392.3A CN111428740A (en) | 2020-02-28 | 2020-02-28 | Detection method and device for network-shot photo, computer equipment and storage medium |
| PCT/CN2021/070734 WO2021169625A1 (en) | 2020-02-28 | 2021-01-08 | Method and apparatus for detecting reproduced network photograph, computer device, and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010127392.3A CN111428740A (en) | 2020-02-28 | 2020-02-28 | Detection method and device for network-shot photo, computer equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111428740A true CN111428740A (en) | 2020-07-17 |
Family
ID=71547808
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010127392.3A Pending CN111428740A (en) | 2020-02-28 | 2020-02-28 | Detection method and device for network-shot photo, computer equipment and storage medium |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN111428740A (en) |
| WO (1) | WO2021169625A1 (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112364856A (en) * | 2020-11-13 | 2021-02-12 | 润联软件系统(深圳)有限公司 | Method and device for identifying copied image, computer equipment and storage medium |
| WO2021169625A1 (en) * | 2020-02-28 | 2021-09-02 | 深圳壹账通智能科技有限公司 | Method and apparatus for detecting reproduced network photograph, computer device, and storage medium |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113222952B (en) * | 2021-05-20 | 2022-05-24 | 蚂蚁胜信(上海)信息技术有限公司 | Method and device for identifying copied image |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102721952A (en) * | 2012-05-16 | 2012-10-10 | 西安电子科技大学 | Method for steadily classifying ground moving target based on super-resolution Doppler spectrum |
| CN107154024A (en) * | 2017-05-19 | 2017-09-12 | 南京理工大学 | Dimension self-adaption method for tracking target based on depth characteristic core correlation filter |
| CN108549836A (en) * | 2018-03-09 | 2018-09-18 | 通号通信信息集团有限公司 | Reproduction detection method, device, equipment and the readable storage medium storing program for executing of photo |
| CN109472768A (en) * | 2018-09-19 | 2019-03-15 | 上海泛洲信息科技有限公司 | A method of in kind and non-flat image in kind is distinguished using spectrum analysis |
| CN109886275A (en) * | 2019-01-16 | 2019-06-14 | 深圳壹账通智能科技有限公司 | Reproduction image-recognizing method, device, computer equipment and storage medium |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105957080A (en) * | 2016-04-28 | 2016-09-21 | 王超维 | Screen-photographed identity card photo identification method based on frequency domain |
| CN111428740A (en) * | 2020-02-28 | 2020-07-17 | 深圳壹账通智能科技有限公司 | Detection method and device for network-shot photo, computer equipment and storage medium |
-
2020
- 2020-02-28 CN CN202010127392.3A patent/CN111428740A/en active Pending
-
2021
- 2021-01-08 WO PCT/CN2021/070734 patent/WO2021169625A1/en active Application Filing
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102721952A (en) * | 2012-05-16 | 2012-10-10 | 西安电子科技大学 | Method for steadily classifying ground moving target based on super-resolution Doppler spectrum |
| CN107154024A (en) * | 2017-05-19 | 2017-09-12 | 南京理工大学 | Dimension self-adaption method for tracking target based on depth characteristic core correlation filter |
| CN108549836A (en) * | 2018-03-09 | 2018-09-18 | 通号通信信息集团有限公司 | Reproduction detection method, device, equipment and the readable storage medium storing program for executing of photo |
| CN109472768A (en) * | 2018-09-19 | 2019-03-15 | 上海泛洲信息科技有限公司 | A method of in kind and non-flat image in kind is distinguished using spectrum analysis |
| CN109886275A (en) * | 2019-01-16 | 2019-06-14 | 深圳壹账通智能科技有限公司 | Reproduction image-recognizing method, device, computer equipment and storage medium |
Non-Patent Citations (1)
| Title |
|---|
| 曾国良，丁宗豪: "频谱分析和DSP频谱分析仪", 电视技术, no. 03, 17 March 1994 (1994-03-17), pages 11 - 18 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021169625A1 (en) * | 2020-02-28 | 2021-09-02 | 深圳壹账通智能科技有限公司 | Method and apparatus for detecting reproduced network photograph, computer device, and storage medium |
| CN112364856A (en) * | 2020-11-13 | 2021-02-12 | 润联软件系统(深圳)有限公司 | Method and device for identifying copied image, computer equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2021169625A1 (en) | 2021-09-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11609968B2 (en) | Image recognition method, apparatus, electronic device and storage medium | |
| US8908911B2 (en) | Redundant detection filtering | |
| CN106560840B (en) | A kind of image information identifying processing method and device | |
| CN109116129B (en) | Terminal detection method, detection device, system and storage medium | |
| CN109948420B (en) | Face comparison method and device and terminal equipment | |
| CN111275685A (en) | Method, device, equipment and medium for identifying copied image of identity document | |
| CN111626163B (en) | Human face living body detection method and device and computer equipment | |
| CN111428740A (en) | Detection method and device for network-shot photo, computer equipment and storage medium | |
| CN106651797B (en) | Method and device for determining effective area of signal lamp | |
| CN115131714A (en) | Intelligent detection and analysis method and system for video image | |
| CN111259915A (en) | Method, device, equipment and medium for recognizing copied image | |
| WO2023025010A1 (en) | Stroboscopic banding information recognition method and apparatus, and electronic device | |
| CN114298902A (en) | Image alignment method and device, electronic equipment and storage medium | |
| CN111915541A (en) | Image enhancement processing method, device, equipment and medium based on artificial intelligence | |
| CN104202448A (en) | System and method for solving shooting brightness unevenness of mobile terminal camera | |
| CN111127358A (en) | Image processing method, device and storage medium | |
| CN111881740B (en) | Face recognition method, device, electronic equipment and medium | |
| CN110689478B (en) | Image stylization processing method and device, electronic equipment and readable medium | |
| US11348254B2 (en) | Visual search method, computer device, and storage medium | |
| CN112052702A (en) | Method and device for identifying two-dimensional code | |
| CN114119964A (en) | Network training method and device, and target detection method and device | |
| CN112633200A (en) | Human face image comparison method, device, equipment and medium based on artificial intelligence | |
| CN108270973B (en) | Photographing processing method, mobile terminal and computer readable storage medium | |
| CN112258556B (en) | Tracking method and device for designated area in video, readable medium and electronic equipment | |
| CN112819859B (en) | Multi-target tracking method and device applied to intelligent security |
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 |