Steganalysis method, device, equipment, medium and product
By performing three-primary color channel image enhancement processing and feature analysis on color images, combining attention mechanism and multi-scale information extraction, the problem that traditional steganography analysis methods are difficult to deal with color images is solved, and more efficient steganography analysis performance and accuracy are achieved.
Patent Information
- Application Number
- CN202411745071.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-29
- Publication Date
- 2025-03-21
AI Technical Summary
The traditional steganography analysis method mainly targets grayscale images, making it difficult to effectively process color images, and is unable to adaptively learn important features of each channel, resulting in the network losing important features during steganography analysis, and insufficient representation and generalization capabilities.
By obtaining the initially recognized image, image enhancement processing is performed on the image based on its three primary color channels, target feature information is extracted, and these feature information is characterized to obtain steganography analysis results. This method combines attention mechanism and multi-scale information extraction, which can more effectively analyze different channel features of color images.
Stereograph analysis for color images is realized, detection performance and accuracy are improved, steganography features in different channels can be better integrated, and the representation and generalization capabilities of the model are improved.
Smart Images
Figure CN119671836A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of image processing technology, and in particular to a steganalysis method, device, equipment, medium and product. Background Art
[0002] In recent years, steganalysis has mainly focused on grayscale images. This is because grayscale images have simpler features and structures than color images, and the related technologies are relatively mature.
[0003] Traditional steganalysis methods are usually designed for grayscale images, and identify steganographic images by extracting low-level features such as image texture and statistical features. However, with the popularity of color images in practical applications, traditional grayscale image steganalysis methods can no longer adapt well to the processing needs of color images. Therefore, it is urgent to form a steganalysis method that can process color images. Summary of the invention
[0004] The present invention provides a steganalysis method, device, equipment, medium and product to realize steganalysis for color images.
[0005] In a first aspect, the present disclosure provides a steganalysis method, comprising:
[0006] Obtaining an initial image to be recognized;
[0007] Based on the three primary color channels of the initial image to be identified, performing image enhancement processing on the image to be identified to obtain target feature information;
[0008] Perform feature analysis on the target feature information to obtain a steganalysis result.
[0009] In a second aspect, the present disclosure provides a steganalysis device, comprising:
[0010] An image acquisition module, used to acquire an initial image to be recognized;
[0011] An image processing module, used for performing image enhancement processing on the image to be identified based on the three primary color channels of the initial image to be identified to obtain target feature information;
[0012] The steganalysis module is used to perform feature analysis on the target feature information to obtain a steganalysis result.
[0013] In a third aspect, an embodiment of the present disclosure provides an electronic device, including:
[0014] at least one processor; and
[0015] a memory communicatively connected to at least one processor; wherein,
[0016] The memory stores a computer program that can be executed by at least one processor, and the computer program is executed by at least one processor so that the at least one processor can execute a steganalysis method provided by the first aspect embodiment above.
[0017] In a fourth aspect, an embodiment of the present disclosure provides a computer-readable storage medium, which stores computer instructions, and the computer instructions are used to enable a processor to implement a steganalysis method provided by the embodiment of the first aspect above when executed.
[0018] In a fifth aspect, an embodiment of the present disclosure provides a computer program product, which includes a computer program. When the computer program is executed by a processor, it implements a steganalysis method provided by the embodiment of the first aspect.
[0019] A steganalysis method, device, equipment, medium and product of an embodiment of the present invention comprises obtaining an initial image to be identified; performing image enhancement processing on the image to be identified based on the three primary color channels of the initial image to be identified to obtain target feature information; performing feature analysis on the target feature information to obtain a steganalysis result. The above technical solution implements steganalysis for color images based on the three primary color enhancement and recognition processing of the image.
[0020] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS
[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0022] Figure 1 is a flow chart of a steganalysis method provided in Embodiment 1 of the present invention;
[0023] Figure 2 is a flow chart of a steganalysis method provided by Embodiment 2 of the present invention;
[0024] Figure 3 It is a structural schematic diagram of a steganalysis device provided by Embodiment 3 of the present invention;
[0025] Figure 4 It is a structural schematic diagram of an electronic device provided in Embodiment 4 of the present invention. DETAILED DESCRIPTION
[0026] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.
[0027] It should be noted that the terms "first", "second" and "target" in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.
[0028] Embodiment 1
[0029] Image steganography is a technical means of embedding information into images for covert communication. Since images have more redundant space than other carriers such as audio, information is more difficult to be discovered. Image steganography has been widely used. Image steganalysis is to detect the existence of steganographic signals and classify images into carrier images without embedded information and secret images with embedded information.
[0030] Most of the existing steganalysis methods are aimed at single-channel grayscale images. Most of the steganalysis of color images uses a single-channel grayscale steganalysis algorithm to analyze the three channels separately. When using the grayscale image single-channel steganalysis method to extract the features of each channel for analysis, the correlation between channels is ignored, and the important features of each channel cannot be adaptively learned, making it impossible for the network to better focus on the features that are more important for steganalysis. In addition, due to the fixed size of the convolution kernel, the feature extraction field of view is limited, focusing on local features while ignoring global features. The existing methods lack feature analysis from local details to global features, which may result in the loss of important features related to steganalysis during network analysis, resulting in insufficient representation and generalization capabilities of the steganalysis network.
[0031] There are several challenges in steganalysis of color images:
[0032] 1. Increased data complexity: Color images have higher information content and complexity than grayscale images, and need to process more color channels and texture changes.
[0033] 2. Difficulty in feature extraction: Feature extraction in color images is relatively complex, and traditional grayscale image feature extraction methods are often not directly applicable.
[0034] 3. Advances in steganography: With the continuous development of steganography, modern steganography can often hide information better, which puts higher requirements on steganalysis.
[0035] Therefore, in order to solve the above technical problems, this embodiment provides a steganalysis method for color images.
[0036] Figure 1 It is a flow chart of a steganalysis method provided in the first embodiment of the present invention. This embodiment is applicable to the case of performing steganalysis on a color image. The method can be executed by a steganalysis device, and the steganalysis device can be implemented in the form of hardware and / or software.
[0037] like Figure 1 As shown, the method includes:
[0038] S101: Obtain an initial image to be recognized.
[0039] In this embodiment, the initial image to be identified can be understood as the initial image to be steganalyzed.
[0040] Specifically, an initial image to be identified is obtained as input to perform steganalysis on the initial image to be identified, wherein the initial image to be identified is a color image. Different from a single-channel grayscale image, a color image has three channels of red (Red, R), green (Green, G), and blue (Blue, B). The initial image to be identified may be a carrier image or a secret image.
[0041] S102 , based on the three primary color channels of the initial image to be identified, perform image enhancement processing on the image to be identified to obtain target feature information.
[0042] In this embodiment, the target feature information can be understood as feature information extracted after feature enhancement processing is performed on the initial image to be identified.
[0043] Specifically, a color image has channels of three primary colors: red, green, and blue. Therefore, the three primary color channels are called to divide the image to be identified into three color feature images, and enhancement processing is performed on each color feature. The enhanced image features are then spliced to form enhanced target feature information.
[0044] S103, performing feature analysis on the target feature information to obtain a steganalysis result.
[0045] In this embodiment, the steganalysis result can be understood as the result of steganalysis of a color image, including image types and type probability values. Image types include carrier images and secret-carrying images.
[0046] Specifically, the target feature information is input into a specific steganalysis model, and the feature analysis of the target feature information is performed through the model to obtain the steganalysis result, thereby determining whether the image to be identified is a carrier image or a secret image.
[0047] It can be understood that steps S102-S103 in this embodiment are applied to a model based on the attention mechanism and multi-scale information extraction, that is, the acquired initial image to be identified is input into the model, and feature enhancement processing is performed through the attention module and the multi-scale information extraction module in the model, and then type recognition and probability analysis of the feature information before and after enhancement are performed based on a specific neural network sub-model, and finally the steganalysis result output by the model is obtained.
[0048] The embodiment of the present invention provides a steganalysis method, comprising obtaining an initial image to be identified; performing image enhancement processing on the image to be identified based on the three primary color channels of the initial image to be identified to obtain target feature information; and performing feature analysis on the target feature information to obtain a steganalysis result. The above technical solution realizes steganalysis for color images and improves the detection performance and accuracy of color image steganalysis.
[0049] Embodiment 2
[0050] Figure 2 It is a flow chart of a steganalysis method provided in the second embodiment of the present invention. This embodiment is a further optimization of any of the above embodiments and can be applied to the case of performing steganalysis on color images. The method can be executed by a steganalysis device, and the steganalysis device can be implemented in the form of hardware and / or software.
[0051] like Figure 2 As shown, the method includes:
[0052] S201: Obtain an initial image to be recognized.
[0053] S202 , based on the three primary color channels of the image to be identified, perform feature extraction processing on the image to be identified, and obtain three target feature maps corresponding to the three primary color channels.
[0054] In this embodiment, the target feature map can be understood as an image corresponding to the three primary colors formed after the image to be identified is split and enhanced, including a red feature map, a green feature map, and a blue feature map.
[0055] Specifically, based on the RGB three primary color channels of the image to be identified, the three primary color features of the image to be identified are extracted, and for each color feature, image information extraction and feature enhancement are performed to form a target feature map after feature enhancement processing corresponding to the color, forming a total of three target feature maps.
[0056] Optionally, based on the three primary color channels of the image to be identified, feature extraction processing is performed on the image to be identified to obtain three target feature maps corresponding to the three primary color channels, including:
[0057] S2021. Divide the image to be identified into three channel images based on the three primary color channels of the image to be identified.
[0058] In this embodiment, the channel image can be understood as an image directly formed after the image to be identified is split based on the three primary color channels, including a red channel image, a green channel image, and a blue channel image.
[0059] Specifically, the image to be identified is subjected to color image extraction according to the three primary color channels to form a red channel image corresponding to the R channel, a green channel image corresponding to the G channel, and a blue channel image corresponding to the B channel.
[0060] S2022. For each channel image, perform multi-scale information extraction processing on the channel image to obtain a preset number of channel feature maps corresponding to the channel image.
[0061] In this embodiment, the channel feature map can be understood as an image formed after feature extraction is performed on the original channel image.
[0062] Specifically, for each channel image, the channel image is processed by a multi-scale information extraction module, wherein the multi-scale information extraction module is composed of a preset number of convolution kernels, and finally forms a preset number of channel feature maps.
[0063] Exemplarily, the multi-scale information extraction module is composed of three convolution kernels of 3*3, 5*5, and 16*16. Each channel image obtains three different channel feature maps, and three channel images obtain nine channel feature maps.
[0064] S2023. Concatenate a preset number of channel feature maps to obtain a target feature map corresponding to the channel image.
[0065] In this embodiment, for each channel, channel image features in the channel are extracted to form a preset number of channel feature maps, and these channel feature maps are spliced to form a target feature map corresponding to the channel.
[0066] For example, each of the three channels R, G, and B has three corresponding channel feature maps, a total of nine. For each channel, the three channel feature maps in the channel are merged, and the three channels R, G, and B form a total of three target feature maps, namely, the red feature map of R f1 , the green feature map G of the G channel f1 And the blue feature map B of the B channel f1 .
[0067] S203: For each target feature map, perform weighted processing on the target feature map to obtain a weighted feature map.
[0068] In this embodiment, the weighted feature map can be understood as a feature map after the target feature map is assigned a weight, including a red weighted feature map, a green weighted feature map, and a blue weighted feature map.
[0069] Specifically, each target feature map is a feature map of a channel, but the feature maps ultimately need to be spliced into a complete image to be recognized, so the splicing method needs to be determined in advance, that is, the splicing weight. Therefore, for each target feature map, the weight of the target feature map is determined based on the attention selective enhancement (Squeeze-and-Excitation, SE) mechanism, and the target feature map is weighted and assigned a weight value to form a weighted feature map. The three channels R, G, and B form a total of three weighted feature maps.
[0070] Optionally, for each target feature map, weighted processing is performed on the target feature map to obtain a weighted feature map, including:
[0071] S2031. For each target feature map, perform pooling processing on the target feature map to obtain a corresponding feature scalar.
[0072] In this embodiment, each target feature map is input into the SE module, and compression processing is performed based on the SE module, and the target feature map is globally pooled to convert the target feature map into a feature scalar.
[0073] S2032. Perform full connection processing on the feature scalar to obtain a target weight corresponding to the target feature map.
[0074] In this embodiment, the target weight can be understood as the weight of the target feature map, and the target weight includes red weight, green weight and blue weight.
[0075] Specifically, for the feature scalar of each target feature map, excitation processing is performed based on the SE module, and the feature scalar is input into a fully connected network (Fully Convolutional Networks, FCN) to obtain the target weights corresponding to the feature scalar, that is, the target weights corresponding to the target feature map, and a total of three target weights corresponding to the three target feature maps are obtained. Among them, the fully connected network consists of a hidden layer and an activation function, and the activation function is a linear rectified function (Rectified Linear Unit, ReLU).
[0076] S2033. Perform weighted processing on the target feature map based on the target weight to obtain a corresponding weighted feature map.
[0077] In this embodiment, the target weight is multiplied by the corresponding target feature map to obtain a weighted feature map for the target feature map, and a red weighted feature map for the red feature map, a green weighted feature map for the green feature map, and a blue weighted feature map for the blue feature map are obtained.
[0078] S204: Concatenate the three weighted feature maps to obtain target feature information.
[0079] In this embodiment, three weighted feature maps (red weighted feature map, green weighted feature map and blue weighted feature map) of the three channels of R, G and B are spliced to form a large feature map to obtain the target feature information of the entire image.
[0080] S205 , inputting the target feature information into a steganalysis model to obtain a steganalysis result of the target feature information, wherein the steganalysis result includes an image type and a type probability value.
[0081] In this embodiment, the steganalysis model can be understood as a model for steganalysis, which is a neural network model, for example, it can be a wider separation-recombination network (WISERNet) or other neural networks, and this embodiment does not limit this.
[0082] Specifically, the target feature information is input into the WISERNet in the steganalysis model. WISERNet consists of a 72*256*256 convolution layer, a 288*128*128 convolution layer, a 1152*32*32 convolution layer, and four Dense layers. Finally, the steganalysis result is used to determine whether the output image is a carrier image or a secret image. Among them, the steganalysis result includes the image type and type probability value. For example, the probability that the target network to be identified is a carrier image is 80%, and the probability that the target image is a secret image is 20%. It can be seen that the probability of the carrier image far exceeds the probability of the secret image, or the probability value of the carrier image exceeds a specific threshold (for example, 60%). Therefore, the target feature information (initial image to be identified) can be directly determined to be a carrier image.
[0083] It is understandable that the steganalysis model is a pre-trained neural network model. Therefore, before the steganalysis model is put into use, it needs to be iteratively trained. The training steps include:
[0084] a. Obtain multiple carrier images and secret images, and use the carrier images and secret images as training samples.
[0085] In this embodiment, a large number of carrier images and encrypted images are acquired to form an image data set, the images in the data set are shuffled, and the shuffled image data set is used as a training sample set.
[0086] b. Iteratively train the initial model based on the training samples until the convergence conditions are met to form a fully trained steganalysis model.
[0087] In this embodiment, the initial model can be understood as an untrained neural network model. The convergence condition can be understood as a condition used to characterize the completeness of model training, such as the loss function is less than a specific threshold.
[0088] Specifically, each training sample in the training sample set is input into the initial model, and based on the model output result (whether the image is a carrier image or a secret image), and confirming whether the model output result is correct, the steganalysis accuracy for the entire training data set is obtained, and the accuracy = the number of correctly identified steganalysis images / the total number of steganalysis images. If the accuracy is low, the loss function or other model parameters of the model are updated, and the next round of training is carried out, and the iterative cycle is carried out until the accuracy is greater than a specific threshold or the loss function converges to a minimum value, forming a fully trained steganalysis model.
[0089] Among them, the loss function calculation uses cross entropy loss to calculate the loss between the true label and the network output, and reduces the loss between the two through back propagation and gradient descent algorithm. Where N is the number of images, y irepresents the label of sample i, where label 0 represents a carrier image without embedded information, and label 1 represents a cryptanalyst image with embedded information. i It represents the probability that sample i is predicted to be a confidential image.
[0090] It can be understood that the above training steps are applicable to the steganalysis model, and are also applicable to the model represented by steps S202-S205 of this embodiment, which is a new deep learning network structure that introduces the SE attention mechanism and multi-scale information extraction to realize steganalysis for color images. The steganalysis model is a sub-model in this model. The SE attention mechanism and multi-scale information extraction are combined to improve the detection performance and accuracy of color image steganography. By using deep learning technology to perform more effective feature learning and representation of color images, the challenges of color image steganalysis can be better addressed, thereby improving the accuracy and robustness of steganalysis.
[0091] A steganalysis method provided by an embodiment of the present invention includes obtaining an initial image to be identified; based on the three primary color channels of the image to be identified, performing feature extraction processing on the image to be identified to obtain three target feature maps corresponding to the three primary color channels; for each target feature map, performing weighted processing on the target feature map to obtain a weighted feature map; splicing the three weighted feature maps to obtain target feature information; inputting the target feature information into a steganalysis model to obtain a steganalysis result of the target feature information, the steganalysis result including an image type and a type probability value. The above technical solution, based on the combination of SE attention mechanism and multi-scale information, can analyze the correlation between different channels of a color image, help to fuse the steganalysis features in different channels, and improve the accuracy of steganalysis; extract the local details of the image to the global features through convolution kernels of different scales, effectively integrate information of different scales, resolutions or levels into the model, so that the model is not limited by the size of the convolution kernel, and also has the ability to pay attention to global features, so as to improve its performance and generalization ability, enhance the representation ability of the model, and is more practical and effective than the existing steganalysis method for grayscale images.
[0092] Embodiment 3
[0093] Figure 3 Schematic diagram of the structure of a steganalysis device provided by Embodiment 3 of the present invention. Figure 3 As shown, the device comprises:
[0094] An image acquisition module 31 is used to acquire an initial image to be recognized;
[0095] An image processing module 32 is used to perform image enhancement processing on the image to be identified based on the three primary color channels of the initial image to be identified to obtain target feature information;
[0096] The steganalysis module 33 is used to perform feature analysis on the target feature information to obtain a steganalysis result.
[0097] The steganalysis device adopted in the technical solution realizes steganalysis for color images based on the three primary color enhancement and recognition processing of the image.
[0098] Optionally, the image processing module 32 includes:
[0099] A feature extraction unit, configured to perform feature extraction processing on the image to be identified based on the three primary color channels of the image to be identified, so as to obtain three target feature maps corresponding to the three primary color channels;
[0100] A feature weighting unit, used for performing weighted processing on each target feature map to obtain a weighted feature map;
[0101] The feature map concatenation unit is used to concatenate the three weighted feature maps to obtain target feature information.
[0102] Optionally, a feature extraction unit is specifically used to:
[0103] Based on the three primary color channels of the image to be identified, dividing the image to be identified into three channel images;
[0104] For each channel image, perform multi-scale information extraction processing on the channel image to obtain a preset number of channel feature maps corresponding to the channel image;
[0105] The preset number of channel feature maps are spliced together to obtain a target feature map corresponding to the channel image.
[0106] Optionally, a feature weighting unit is used to:
[0107] For each of the target feature maps, a pooling process is performed on the target feature map to obtain a corresponding feature scalar;
[0108] Performing full connection processing on the feature scalar to obtain a target weight corresponding to the target feature map;
[0109] The target feature map is weighted based on the target weight to obtain a corresponding weighted feature map.
[0110] Optionally, the steganalysis module 33 is specifically used for:
[0111] The target feature information is input into a steganalysis model to obtain a steganalysis result of the target feature information, wherein the steganalysis result includes an image type and a type probability value.
[0112] Optionally, the device further includes a model training module, which is used to:
[0113] Acquire a plurality of carrier images and secret-carrying images, and use the carrier images and secret-carrying images as training samples;
[0114] The initial model is iteratively trained based on the training samples until a convergence condition is met, thereby forming a fully trained steganalysis model.
[0115] The steganalysis device provided in the embodiment of the present invention can execute the steganalysis method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.
[0116] Embodiment 4
[0117] Figure 4 : is a structural schematic diagram of an electronic device provided in Embodiment 4 of the present invention. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workbenches, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.
[0118] like Figure 4 As shown, the electronic device 40 includes at least one processor 41, and a memory connected to the at least one processor 41, such as a read-only memory (ROM) 42, a random access memory (RAM) 43, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 41 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 42 or the computer program loaded from the storage unit 48 to the random access memory (RAM) 43. In the RAM 43, various programs and data required for the operation of the electronic device 40 can also be stored. The processor 41, the ROM 42, and the RAM 43 are connected to each other through a bus 44. An input / output (I / O) interface 45 is also connected to the bus 44.
[0119] A number of components in the electronic device 40 are connected to the I / O interface 45, including: an input unit 46, such as a keyboard, a mouse, etc.; an output unit 47, such as various types of displays, speakers, etc.; a storage unit 48, such as a disk, an optical disk, etc.; and a communication unit 49, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 49 allows the electronic device 40 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.
[0120] The processor 41 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 41 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 41 executes the various methods and processes described above, such as the steganalysis method.
[0121] In some embodiments, the steganalysis method may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 48. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 40 via the ROM 42 and / or the communication unit 49. When the computer program is loaded into the RAM 43 and executed by the processor 41, one or more steps of the steganalysis method described above may be performed. Alternatively, in other embodiments, the processor 41 may be configured to perform the steganalysis method in any other suitable manner (e.g., by means of firmware).
[0122] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.
[0123] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.
[0124] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
[0125] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).
[0126] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.
[0127] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.
[0128] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.
[0129] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims
1. A steganalysis method, characterized in that: include: Obtaining an initial image to be recognized; Based on the three primary color channels of the initial image to be identified, performing image enhancement processing on the image to be identified to obtain target feature information; Perform feature analysis on the target feature information to obtain a steganalysis result.
2. The method according to claim 1, characterized in that The step of performing image enhancement processing on the image to be identified based on the three primary color channels of the initial image to be identified to obtain target feature information includes: Based on the three primary color channels of the image to be identified, feature extraction processing is performed on the image to be identified to obtain three target feature maps corresponding to the three primary color channels; For each of the target feature maps, weighted processing is performed on the target feature map to obtain a weighted feature map; The three weighted feature maps are concatenated to obtain target feature information.
3. The method according to claim 2, characterized in that The method of performing feature extraction processing on the image to be identified based on the three primary color channels of the image to be identified to obtain three target feature maps corresponding to the three primary color channels includes: Based on the three primary color channels of the image to be identified, dividing the image to be identified into three channel images; For each channel image, perform multi-scale information extraction processing on the channel image to obtain a preset number of channel feature maps corresponding to the channel image; The preset number of channel feature maps are spliced together to obtain a target feature map corresponding to the channel image.
4. The method according to claim 2, characterized in that: The step of performing weighted processing on each target feature map to obtain a weighted feature map includes: For each of the target feature maps, a pooling process is performed on the target feature map to obtain a corresponding feature scalar; Performing full connection processing on the feature scalar to obtain a target weight corresponding to the target feature map; The target feature map is weighted based on the target weight to obtain a corresponding weighted feature map.
5. The method according to claim 1, characterized in that The step of performing feature analysis on the target feature information to obtain a steganalysis result includes: The target feature information is input into a steganalysis model to obtain a steganalysis result of the target feature information, wherein the steganalysis result includes an image type and a type probability value.
6. The method according to claim 5, characterized in that Also includes: Acquire a plurality of carrier images and secret-carrying images, and use the carrier images and secret-carrying images as training samples; The initial model is iteratively trained based on the training samples until a convergence condition is met, thereby forming a fully trained steganalysis model.
7. A steganalysis device, characterized in that: include: An image acquisition module, used to acquire an initial image to be recognized; An image processing module, used for performing image enhancement processing on the image to be identified based on the three primary color channels of the initial image to be identified to obtain target feature information; The steganalysis module is used to perform feature analysis on the target feature information to obtain a steganalysis result.
8. An electronic device, characterized in that: include: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can perform a steganalysis method according to any one of claims 1 to 6.
9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement a steganalysis method according to any one of claims 1 to 6 when executed.
10. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the computer program implements a steganalysis method according to any one of claims 1 to 6.