CN111476865B - Image protection method for image recognition based on deep learning neural network - Google Patents
Image protection method for image recognition based on deep learning neural network Download PDFInfo
- Publication number
- CN111476865B CN111476865B CN202010211679.4A CN202010211679A CN111476865B CN 111476865 B CN111476865 B CN 111476865B CN 202010211679 A CN202010211679 A CN 202010211679A CN 111476865 B CN111476865 B CN 111476865B
- Authority
- CN
- China
- Prior art keywords
- image
- photo frame
- deep learning
- target
- size
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T11/00—2D [Two Dimensional] image generation
- G06T11/60—Editing figures and text; Combining figures or text
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- 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
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/24—Aligning, centring, orientation detection or correction of the image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/32—Normalisation of the pattern dimensions
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N1/00—Scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission; Details thereof
- H04N1/44—Secrecy systems
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Multimedia (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Health & Medical Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Signal Processing (AREA)
- Image Analysis (AREA)
Abstract
The invention provides an image protection method for image recognition by a deep learning neural network, and relates to the technical field of intelligent information security. The method comprises the steps of protecting with target misleading and protecting without target misleading, firstly collecting image data or using the existing data as a training data set, preprocessing a data image in the training data set, and converting an original image into an input format accepted by a deep learning image classifier model; determining the size of the photo frame according to the size of the original image; finally, under the determined photo frame size, generating a photo frame by utilizing training data, and placing the preprocessed image into the photo frame to realize encryption of the image. The encrypted image user obtained by the method can easily distinguish the image content, and has better user experience.
Description
Technical Field
The invention relates to the technical field of intelligent information security, in particular to an image protection method for image recognition by a deep learning neural network.
Background
With the continuous development of science and technology, the electronic computer technology is increasingly widely applied, the arrival of the Internet age brings about the distance between people, and a large amount of multimedia information can be easily and rapidly transmitted in a network, so that communication and exchange among people are facilitated. People enjoy the convenience of the internet and their personal privacy information is also under test. In particular, in the current data age, the problem is more remarkable, and the influence scope is wider. The development of artificial intelligence makes the life of people become more transparent, and privacy infringement means are more concealed and technically realized, thereby invisibly threatening the data security of people. Especially, people like to send photo videos related to themselves by various social media or store photos by cloud disk software and the like, the photos often contain identity information related to the people themselves or other sensitive information and the like, and the confidentiality of information transmission is important only when people want to see the people or people with specific identities.
Meanwhile, the research of deep learning promotes the development of artificial intelligence, and is applied to more and more fields, for example, the image can be easily identified and detected, and the manpower is greatly liberated. As a result, lawbreakers can easily screen out key information from the stolen mass data, thereby implementing targeted fraudulent theft and the like.
The conventional coping method is to encrypt information, and when viewing the information, a user needs to decrypt the encrypted information to see the original information. For the image photo, the encryption and decryption of the user are complicated, and particularly for the mobile terminal user, the man-machine interaction is not friendly enough. Meanwhile, the user often has a large number of images, on the premise of not decrypting, the user can hardly distinguish the images to be searched, and decrypting all the images is too time-consuming and labor-consuming, so that a protection method is needed, the current image photo can be conveniently protected, the content of the current image can be easily checked, and the user experience is improved.
Disclosure of Invention
The invention aims to solve the technical problem of providing an image protection method for image recognition by a deep learning neural network aiming at the defects of the prior art, and the image protection method aims to attack the image recognition by using a deep learning model.
In order to solve the technical problems, the invention adopts the following technical scheme: an image protection method for image recognition facing a deep learning neural network comprises two protection types of target misleading and protection without target misleading, and comprises the following steps:
step 1, acquiring a certain number of images, and labeling a data set as a training data set aiming at the content of the images, or using the existing data set as the training data set;
step 2, preprocessing a data image in the training data set, and converting an original image in the training data set into an input format accepted by a deep learning image classifier model;
the preprocessing comprises shape adjustment, size adjustment, data range adjustment and image rotation of an image; the shape of the image is adjusted to adjust the length-width ratio of the input image, so that the input data accords with the input requirement of the classifier model; the size adjustment of the image refers to the zooming and cutting of the image, and the length and the width of the image are enlarged or reduced, or the image meeting the input size requirement of the deep learning image classifier model is obtained by cutting; the data range adjustment of the image refers to adjustment of the pixel value range of the image data, and the pixel value range is normalized to a range suitable for the input of a deep learning image classifier model; the rotation of the image means that the image is rotated by a certain angle, so that the complexity and diversity of input data are increased, and the data quality is improved;
the deep learning image classifier model outputs class labels to which images belong and probability of the class labels;
step 3, determining the size of the photo frame according to the size of the original image, namely the pixel width of each side of the photo frame;
step 4, under the determined size of the photo frame, generating the photo frame by utilizing training data, and placing the preprocessed image into the photo frame to encrypt the image;
setting an encrypted image x 'to be synthesized by an original image and a photo frame, wherein the photo frame is represented by delta, and x' =x (1-m) +delta-m, wherein m represents a mask for controlling an image processing area, the size of m is consistent with the size of the encrypted image, the size of the m is 1 at the position corresponding to the photo frame, and the rest positions are 0, so that the original image and the photo frame are synthesized into the encrypted image; the symbol "·" represents pixel-by-pixel multiplication, i.e. multiplication of pixel values at corresponding positions, and the obtained encrypted image has one more frame than the original image, i.e. the central area of the image is the original image and the surrounding is the photo frame;
the photo frame generated by training data is obtained by gradient descent of an optimization problem, and is obtained by randomly sampling from normal distribution at the beginning of solving, namely the photo frame is initialized to be a random number, and then is gradually solved by back propagation; for protection without target induction, a photo frame generated by training data is solved by the following optimization problem:
min(-loss(f(x·(1-m)+δ·m)，y))
where loss () is a loss function, f () is a deep learning image classifier model, y is an original label of an image, i.e., a class of the image, and f (x) =y=argmax (logits (x)));
the protection without target induction means that for any image input to the deep learning image classifier model, the output image prediction category is no longer the correct category, i.e. f (x') +.y;
for image protection with target induction, a photo frame generated by training data is solved by the following optimization problem:
min loss(f(x·(1-m)+δ·m)，y target )
wherein y is target A target tag to be induced;
for the protection finger with target misguidance, regardless of the category of the original image, the deep learning image classifier model predicts the input encrypted image as a specific category, so that the model considers the input image as non-critical information.
The beneficial effects of adopting above-mentioned technical scheme to produce lie in: the image protection method for image recognition facing the deep learning neural network provided by the invention can be directly applied only by generating the photo frame for encryption once, greatly reduces the calculated amount and calculation time, has better practicability, can be completed by simply splicing images when the photo frame is obtained and occupies little calculated amount, and decryption can be completed by removing the photo frame. Meanwhile, the encrypted image user obtained by the method can easily distinguish the image content, and has better user experience.
Drawings
Fig. 1 is a flowchart of an image protection method for performing image recognition on a deep learning neural network according to an embodiment of the present invention;
fig. 2 is a data preprocessing flow chart of an image protection method for image recognition by a deep learning neural network according to an embodiment of the present invention;
fig. 3 is a schematic diagram of an encrypted image according to an embodiment of the present invention, where (a) is an original image, (b) is a photo frame, and (c) is an encrypted image.
Detailed Description
The following describes in further detail the embodiments of the present invention with reference to the drawings and examples. The following examples are illustrative of the invention and are not intended to limit the scope of the invention.
In this embodiment, taking a deep learning image classifier model resnet model as an example, the image protection method for image recognition by the deep learning neural network of the present invention is adopted to protect an image which is subjected to image recognition by the model.
In this embodiment, it is assumed that the state parameters of the model during operation can be obtained, that is, the structure of the resnet model is known, and the parameters of each layer of the resnet model can be obtained, so that when the resnet model predicts the picture, the output of each layer of the model can be deduced, and finally the input type and probability thereof can be deduced.
In this embodiment, the selected resnet model is denoted by f, and for any preprocessed image x, the resnet model performs a classification task to predict the correct class y of the image x, i.e., f (x) =y. This process can be further expressed as: the second last layer of the event is a logits layer, the output is a 1000-dimensional vector, each dimension number corresponds to logits of a corresponding category, the 1000-dimensional vector is used as input of the softmax layer of the last layer to obtain a new 1000-dimensional vector, the probability of the new 1000-dimensional vector corresponding to each category is obtained, and finally the resnet model outputs the category with the highest probability as a prediction result, namely f (x) =y=argmax (logits (x)).
An image protection method for image recognition facing a deep learning neural network, as shown in fig. 1, comprises two protection types of target misleading and protection without target misleading, and comprises the following steps:
step 1, acquiring a certain number of images, and labeling a data set as a training data set aiming at the content of the images, or using the existing data set as the training data set;
in this embodiment, the training data set may be manually obtained by photographing in real life and manually constructing by a web crawler, and the selected picture needs to conform to the classification field of the resnet model.
The training data set may also be obtained from a resnet model, and a representative image conforming to the model is generated as the data set using the known structural parameters of the model.
The training data can be obtained by directly utilizing the existing network data set, and for the resnet model, the training data set is an ImageNet data set, so far, more than twenty thousands of types and more than ten millions of images exist in ImageNet, and a proper number of data sets can be selected as the training set.
Step 2, preprocessing a data image in the training data set, as shown in fig. 2, converting an original image into an input format acceptable by a deep learning image classifier model;
the preprocessing comprises shape adjustment, size adjustment, data range adjustment and image rotation of an image;
the shape of the image is adjusted to adjust the length-width ratio of the input image, so that the input data accords with the input requirement of the classifier model; for example, the aspect ratio of the original image is 4:3, an adjustment is required to make the aspect ratio of the image 1 by using interpolation function or the like: 1. the size adjustment of the image refers to the zooming and cutting of the image, and the length and the width of the image are enlarged or reduced, or the image meeting the input size requirement of the deep learning image classifier model is obtained by cutting; for example, the length and width pixels of the original image are 300, and the size of the image can be adjusted to 224 by using interpolation methods such as bilinear interpolation, cubic spline function and the like, or the image is directly cut at the center of the image, so that an image with the size of 224 x 224 is obtained, and the size accords with the input of a resnet model.
The data range adjustment of the image refers to adjustment of the pixel value range of the image data, and the pixel value range is normalized to a range suitable for the input of a deep learning image classifier model; for example, after the original image is digitized, the pixel value interval is [0,255], normalization is required, and the pixel value range is adjusted to [ -1,1], that is, the input interval of the resnet. Normalization can be performed using a normal distribution, with the mean of the three RGB channels (0.4914,0.4822,0.4465) and the variance (0.2023,0.1994,0.2010).
The rotation of the image means that the image is rotated by a certain angle, for example, the rotation angle is randomly selected from [ -15,15], 15 means that the image is rotated rightwards by 15 degrees, 15 means that the image is rotated leftwards by 15 degrees, and the random rotation of the image can increase the data quantity on one hand, increase the image recognition complexity on the other hand, improve the data quality and enhance the model generalization.
The deep learning image classifier model outputs class labels to which images belong and probability of the class labels;
step 3, determining the size of the photo frame according to the size of the original image, namely the pixel width of each side of the photo frame;
the protection method only needs the photo frame to realize the protection of the image easily, and the bigger the photo frame is, the higher the encryption degree of the image is, and meanwhile, the bigger the proportion of the photo frame to the image is. In this embodiment, the input image size of the resnet model is 224×224, and the width of the boundary pixel area where the photo frame is located is 5, i.e. the photo frame is within 5 pixel ranges around the original image, and the photo frame occupies 8.36% of the area of the encrypted image. When the width of the boundary pixel area where the photo frame is positioned is 6, the photo frame occupies 9.91% of the original image area.
Step 4, under the determined size of the photo frame, generating the photo frame by using training data, and placing the image into the photo frame to encrypt the image;
setting an encrypted image x 'to be synthesized by an original image and a photo frame, wherein the photo frame is represented by delta, and x' =x (1-m) +delta-m, wherein m represents a mask for controlling an image processing area, the size of m is consistent with the size of the encrypted image, the size of the m is 1 at the position corresponding to the photo frame, and the rest positions are 0, so that the original image and the photo frame are synthesized into the encrypted image; the symbol "·" represents pixel-by-pixel multiplication, i.e. multiplication of pixel values at corresponding positions, and the obtained encrypted image has one more frame than the original image, i.e. the central area of the image is the original image and the surrounding is the photo frame; in this embodiment, the resulting encrypted image is shown in fig. 3.
The photo frame generated by training data is obtained by gradient descent of an optimization problem, and is obtained by randomly sampling from normal distribution at the beginning of solving, namely the photo frame is initialized to be a random number, and then is gradually solved by back propagation;
for protection without target induction, a photo frame generated by training data is solved by the following optimization problem:
min(-loss(f(x·(1-m)+δ·m)，y))
where loss () is a loss function, f () is a deep learning image classifier model, y is an original label of an image, i.e., a class of the image, and f (x) =y=argmax (logits (x))); in this embodiment, the loss function is a cross entropy function.
The protection without target induction means that for any image input to the deep learning image classifier model, the output image prediction category is no longer the correct category, i.e. f (x') +.y; for example, for the resnet model originally predicted as class 1, any number from 0 to 999 other than 1 can be predicted after encryption.
The optimization problem increases the loss function value of the encrypted image and the original label y;
for image protection with target induction, a photo frame generated by training data is solved by the following optimization problem:
min loss(f(x·(1-m)+δ·m)，y target )
wherein y is target A target tag to be induced;
for the protection finger with target misleading, disregarding the category of the original image, enabling the deep learning image classifier model to predict the input encrypted image as a specific category, so that the model considers the input image as non-critical information; for example, selecting the target class 859, after encryption, the image is considered as the target class, regardless of the correct class to which the previous image belongs, which makes it unimportant to the sender.
The optimization problem reduces the loss function value between the encrypted image and the target label, so that the encrypted image is more easily divided into target categories by mistake.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the 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 scheme described in the foregoing embodiments can be modified or some or all of the technical features thereof can be replaced with equivalents; such modifications and substitutions do not depart from the spirit of the corresponding technical solutions, which are defined by the scope of the appended claims.
Claims (2)
1. An image protection method for image recognition facing a deep learning neural network is characterized by comprising the following steps of: the method comprises target misleading protection and target misleading-free protection, wherein the target misleading-free protection refers to that for any image input into a deep learning image classifier model, the output image prediction category is no longer the correct category, namely f (x')noteqy;
the protection finger with target misleading does not consider the category of the original image, so that the deep learning image classifier model predicts the input encrypted image into a specific category, and the model considers the input image as non-critical information;
the method comprises the following steps:
step 1, acquiring a certain number of images, and labeling a data set as a training data set aiming at the content of the images, or using the existing data set as the training data set;
step 2, preprocessing a data image in the training data set, and converting an original image in the training data set into an input format accepted by a deep learning image classifier model;
step 3, determining the size of the photo frame according to the size of the original image, namely the pixel width of each side of the photo frame;
step 4, under the determined size of the photo frame, generating the photo frame by utilizing training data, and placing the preprocessed image into the photo frame to encrypt the image;
setting an encrypted image x 'to be synthesized by an original image and a photo frame, wherein the photo frame is represented by delta, and x' =x (1-m) +delta-m, wherein m represents a mask for controlling an image processing area, the size of m is consistent with the size of the encrypted image, the size of the m is 1 at the position corresponding to the photo frame, and the rest positions are 0, so that the original image and the photo frame are synthesized into the encrypted image; the symbol "·" represents pixel-by-pixel multiplication, i.e. multiplication of pixel values at corresponding positions, and the obtained encrypted image has one more frame than the original image, i.e. the central area of the image is the original image and the surrounding is the photo frame;
the photo frame generated by training data is obtained by gradient descent of an optimization problem, and is obtained by randomly sampling from normal distribution at the beginning of solving, namely the photo frame is initialized to be a random number, and then is gradually solved by back propagation;
for protection without target misleading, a photo frame generated by training data is solved through the following optimization problem:
min(-loss(f(x·(1-m)+δ·m)，y))
where loss () is a loss function, f () is a deep learning image classifier model, y is an original label of an image, i.e., a class of the image, and f (x) =y=argmax (logits (x)));
for image protection with target misleading, a photo frame generated by training data is solved through the following optimization problem:
min loss(f(x·(1-m)+δ·m)，y target )
wherein y is target Is the target tag to be misled.
2. The image protection method for image recognition by deep learning neural network according to claim 1, wherein the method comprises the following steps: step 2, preprocessing comprises shape adjustment, size adjustment, data range adjustment and image rotation of an image; the shape of the image is adjusted to adjust the length-width ratio of the input image, so that the input data accords with the input requirement of the classifier model; the size adjustment of the image refers to the zooming and cutting of the image, and the length and the width of the image are enlarged or reduced, or the image meeting the input size requirement of the deep learning image classifier model is obtained by cutting; the data range adjustment of the image refers to adjustment of the pixel value range of the image data, and the pixel value range is normalized to a range suitable for the input of a deep learning image classifier model; the rotation of the image means that the image is rotated by a certain angle, so that the complexity and diversity of input data are increased, and the data quality is improved;
and the deep learning image classifier model outputs class labels to which the images belong and probability of the class labels.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010211679.4A CN111476865B (en) | 2020-03-24 | 2020-03-24 | Image protection method for image recognition based on deep learning neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010211679.4A CN111476865B (en) | 2020-03-24 | 2020-03-24 | Image protection method for image recognition based on deep learning neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111476865A CN111476865A (en) | 2020-07-31 |
| CN111476865B true CN111476865B (en) | 2023-07-07 |
Family
ID=71747713
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010211679.4A Active CN111476865B (en) | 2020-03-24 | 2020-03-24 | Image protection method for image recognition based on deep learning neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111476865B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2022152153A1 (en) * | 2021-01-18 | 2022-07-21 | 北京灵汐科技有限公司 | Image processing method and device, key generation method and device, training method, and computer readable medium |
| CN113808044B (en) * | 2021-09-17 | 2022-11-01 | 北京百度网讯科技有限公司 | Encryption mask determining method, device, equipment and storage medium |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107729765A (en) * | 2017-09-30 | 2018-02-23 | 广东欧珀移动通信有限公司 | Management method, device, storage medium and the electronic equipment of multi-medium data |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060170669A1 (en) * | 2002-08-12 | 2006-08-03 | Walker Jay S | Digital picture frame and method for editing |
| CN1808969A (en) * | 2006-01-16 | 2006-07-26 | 钱丽娜 | Password transmission apparatus and method for weak authentication |
| CN107770406B (en) * | 2017-10-26 | 2019-06-21 | 北京邮电大学 | Image encryption method and device based on the transformation of multi-parameter fractional order and semi-tensor product |
-
2020
- 2020-03-24 CN CN202010211679.4A patent/CN111476865B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107729765A (en) * | 2017-09-30 | 2018-02-23 | 广东欧珀移动通信有限公司 | Management method, device, storage medium and the electronic equipment of multi-medium data |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111476865A (en) | 2020-07-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Wan et al. | A comprehensive survey on robust image watermarking | |
| Muhammad et al. | Image steganography using uncorrelated color space and its application for security of visual contents in online social networks | |
| CN111476865B (en) | Image protection method for image recognition based on deep learning neural network | |
| CN115378574B (en) | Lightweight dynamic image data encryption method and system | |
| US20210342967A1 (en) | Method for securing image and electronic device performing same | |
| CA3184510A1 (en) | Image distribution using composite re-encrypted images | |
| Li et al. | Data hiding in iris image for privacy protection | |
| CN110110535A (en) | A kind of low distortion steganography method based on picture element matrix | |
| Liu et al. | Unauthorized AI cannot recognize me: Reversible adversarial example | |
| Prasad et al. | Logistic map-based image steganography scheme using combined LSB and PVD for security enhancement | |
| Ye et al. | PRA-TPE: Perfectly recoverable approximate thumbnail-preserving image encryption | |
| Cui et al. | Multitask identity-aware image steganography via minimax optimization | |
| CN107169911A (en) | A kind of image encryption method converted based on Arnold transformation | |
| CN110069907A (en) | Big data source tracing method and system based on digital watermarking | |
| Fadhil et al. | Improved Security of a Deep Learning-Based Steganography System with Imperceptibility Preservation | |
| CN106055632B (en) | Video authentication method based on scene frame fingerprints | |
| Davidson et al. | Locating secret messages in images | |
| CN115952529A (en) | User data processing method, computing device and storage medium | |
| Seem et al. | Artificial neural network, convolutional neural network visualization, and image security | |
| Upadhyay et al. | Protection of Digital Image and Text Information Security Using LSB and Crossover Techniques | |
| Li et al. | Adversarial data hiding with only one pixel | |
| Kumar et al. | A Secure Image Watermarking Scheme Based on DWT, SVD and Arnold Transform | |
| Jeremiah et al. | CCTV Footage De-identification for Privacy Protection: A Comprehensive Survey | |
| Haynes | Using image steganography to establish covert communication channels | |
| Neethu et al. | Securing image posts in social networking sites |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |