Panoramic watermark processing method and device for three-dimensional Gaussian scene, equipment and medium

Abstract

The invention relates to the technical field of three-dimensional Gaussian scene watermarking, in particular to a panoramic watermark processing method and device for a three-dimensional Gaussian scene, equipment and a medium. Determining the parameter sensitivity of each first Gaussian unit; screening a second Gaussian unit of which the corresponding parameter sensitivity is greater than a preset parameter sensitivity threshold from all the first Gaussian units; and embedding the copyright message in the second Gaussian unit in the three-dimensional Gaussian scene, thereby completing the embedding operation of the panoramic watermark. According to the embodiment of the invention, the parameter sensitivity of the first Gaussian unit is estimated through the two-dimensional original image and the first rendered image of different visual angles, so that the geometric structure characteristics of the three-dimensional scene are considered when the watermark is embedded, and the watermark information distortion is avoided.

Description

Technical Field

[0001] This invention relates to the field of 3D Gaussian scene watermarking technology, and more specifically, to a panoramic watermarking processing method, apparatus, equipment, and medium for 3D Gaussian scenes. Background Technology

[0002] 3D Gaussian Splatting (3DGS), a groundbreaking technology in the field of 3D computer vision, revolutionizes traditional 3D scene representation and rendering by constructing explicit geometric and appearance models of scenes using discretized sets of 3D Gaussian units. However, the protection of 3D Gaussian scene assets faces severe challenges: due to the discretized and easily editable nature of 3D Gaussian parameters, related digital assets are highly susceptible to unauthorized alteration, and the original ownership of altered assets is difficult to trace.

[0003] Existing 2D watermarking technology does not take into account the geometric characteristics of 3D scenes. During application, once the perspective is switched or the scene is partially edited, the watermark information is easily distorted, which severely limits its adaptability to 3D Gaussian scene assets. Summary of the Invention

[0004] In view of this, the purpose of this application is to provide a panoramic watermarking method, apparatus, device and medium for three-dimensional Gaussian scenes, which can embed watermarks in three-dimensional Gaussian scenes based on different perspectives and avoid watermark information distortion.

[0005] In a first aspect, embodiments of this application provide a panoramic watermarking method for a three-dimensional Gaussian scene, the method comprising: The process involves acquiring a 3D Gaussian scene to be processed, constructing 2D original images of the 3D Gaussian scene from various first-view perspectives, and first rendered images of the 3D Gaussian scene from various first-view perspectives. Each first rendered image is a 2D image obtained by rendering the first Gaussian unit in the 3D Gaussian scene from the corresponding first-view perspective. Based on the degree of difference between the original 2D image and the first rendered image under each first viewpoint corresponding to each first Gaussian unit, the parameter sensitivity of each first Gaussian unit is determined; the parameter sensitivity is used to quantify the degree of influence of the parameter perturbation introduced by embedding a watermark at the corresponding first Gaussian unit on the geometric integrity and rendering visual consistency of the 3D Gaussian scene. From all first Gaussian units, select second Gaussian units whose corresponding parameter sensitivity is greater than a preset parameter sensitivity threshold; The copyright message is embedded in the second Gaussian unit in the three-dimensional Gaussian scene to complete the embedding operation of the panoramic watermark.

[0006] In one possible implementation, the parameter sensitivity of any first Gaussian element is calculated according to the following steps: For each first viewpoint where the first Gaussian unit is located, the parameter sensitivity matrix of the first Gaussian unit under the first viewpoint is calculated based on the difference between the two-dimensional original image and the first rendered image under the first viewpoint. The parameter sensitivity matrices of the first Gaussian unit under all corresponding first viewpoints are fused to obtain the parameter sensitivity of the first Gaussian unit.

[0007] In one possible implementation, calculating the parameter sensitivity matrix of the first Gaussian unit under the first viewpoint based on the difference value between the original two-dimensional image and the first rendered image under the first viewpoint includes: Substituting the difference value between the two-dimensional original image and the first rendered image under the first viewpoint into the following formula, the parameter sensitivity matrix of the first Gaussian unit under the first viewpoint is obtained. ; in, The parameter sensitivity matrix of the first Gaussian element under the first viewpoint; This is the k-th first-person perspective; Let be the parameter set of the i-th first Gaussian unit; For first-person perspective Below, the difference value between the original 2D image and the first rendered image; Values ​​representing the degree of difference The parameter set of the i-th first Gaussian unit The first-order gradient vector; This is a transpose.

[0008] In one possible implementation, fusing the parameter sensitivity matrices of the first Gaussian unit under all corresponding first viewpoints to obtain the parameter sensitivity of the first Gaussian unit includes: Substituting the parameter sensitivity matrix of the first Gaussian unit under all corresponding first viewpoints into the following formula, the parameter sensitivity of the first Gaussian unit is obtained. ; in, For the parameter sensitivity of the i-th first Gaussian unit, Let i be the number of first-view units containing the i-th first Gaussian unit. The trace of the matrix, The parameter sensitivity matrix of the first Gaussian element under the first viewpoint. For the k-th first-person perspective, Let be the parameter set of the i-th first Gaussian unit.

[0009] In one possible implementation, after embedding the copyright message within a second Gaussian unit in the three-dimensional Gaussian scene, the method further includes: The parameters of each third Gaussian unit in the target 3D Gaussian scene are input into the 3D watermark extraction model to obtain the first copyright extraction message; the target 3D Gaussian scene is the 3D Gaussian scene after embedding the copyright message. And / or, input all the second rendered images corresponding to the target 3D Gaussian scene into the 2D watermark extraction model to obtain the second copyright extraction message; each second rendered image is a 2D image obtained by rendering the fourth Gaussian unit in the target 3D Gaussian scene under the corresponding second viewpoint; Based on the matching degree between the first copyright extraction message and the embedded copyright message, and / or the matching degree between the second copyright extraction message and the embedded copyright message, the watermark verification result corresponding to the target 3D Gaussian scene is determined.

[0010] In one possible implementation, the step of inputting the parameters of each third Gaussian unit in the target 3D Gaussian scene into the 3D watermark extraction model to obtain the first copyright extraction message includes: The parameters of each third Gaussian unit are input into the feature extraction module of the three-dimensional watermark extraction model to obtain the local feature vector corresponding to each third Gaussian unit. The local feature vector corresponding to each third Gaussian unit is used to characterize the correlation between the parameters inside the third Gaussian unit, and / or the correlation between the parameters of the third Gaussian unit and the parameters of its neighboring third Gaussian units. Input the local feature vectors corresponding to all third Gaussian units into the feature aggregation module in the three-dimensional watermark extraction model to obtain the first global feature vector corresponding to the target three-dimensional Gaussian scene; The first global feature vector is input into the first copyright message decoding module in the three-dimensional watermark extraction model to obtain the first copyright extraction message.

[0011] In one possible implementation, the step of inputting all the second rendered images corresponding to the target 3D Gaussian scene into a 2D watermark extraction model to obtain the second copyright extraction message includes: Each second rendered image is input into the multi-scale feature extraction layer of the two-dimensional watermark extraction model to obtain the first image feature map of each second rendered image at each preset scale. For each second rendered image, the first image feature map of the second rendered image at all preset scales is input into the first feature fusion layer of the two-dimensional watermark extraction model to obtain the second image feature vector corresponding to the second rendered image; Input the second image feature vectors corresponding to all the second rendered images into the second feature fusion layer in the two-dimensional watermark extraction model to obtain the second global feature vector corresponding to the target three-dimensional Gaussian scene; The second global feature vector is input into the second copyright message decoding module in the two-dimensional watermark extraction model to obtain the second copyright extraction message.

[0012] Secondly, embodiments of this application also provide a panoramic watermarking device for a three-dimensional Gaussian scene, the device comprising: The acquisition module is used to acquire the three-dimensional Gaussian scene to be processed, to construct the two-dimensional original images of the three-dimensional Gaussian scene in each first viewpoint, and the first rendered images of the three-dimensional Gaussian scene in each first viewpoint; each first rendered image is a two-dimensional image obtained by rendering the first Gaussian unit in the three-dimensional Gaussian scene in the corresponding first viewpoint. The sensitivity determination module is used to determine the parameter sensitivity of each first Gaussian unit based on the degree of difference between the two-dimensional original image and the first rendered image under each first viewpoint corresponding to each first Gaussian unit; the parameter sensitivity is used to quantify the degree of influence of the parameter perturbation introduced by embedding watermark at the corresponding first Gaussian unit on the geometric integrity and rendering visual consistency of the three-dimensional Gaussian scene. The Gaussian cell filtering module is used to filter second Gaussian cells from all first Gaussian cells whose corresponding parameter sensitivity is greater than a preset parameter sensitivity threshold. The watermark embedding module is used to embed copyright information within the second Gaussian unit in the three-dimensional Gaussian scene, thereby completing the panoramic watermark embedding operation.

[0013] Thirdly, embodiments of this application also provide an electronic device, including: a processor, a storage medium, and a bus, wherein the storage medium stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor communicates with the storage medium via the bus, and the processor executes the machine-readable instructions to perform the steps of the panoramic watermarking method for a three-dimensional Gaussian scene as described in any of the first aspects.

[0014] Fourthly, embodiments of this application also provide a computer-readable storage medium storing a computer program, which, when executed by a processor, performs the steps of the panoramic watermarking method for a three-dimensional Gaussian scene as described in any of the first aspects.

[0015] This application provides a method, apparatus, device, and medium for panoramic watermarking of a 3D Gaussian scene. The method includes: acquiring a 3D Gaussian scene to be processed, two-dimensional original images corresponding to the 3D Gaussian scene from various first viewpoints, and first rendered images corresponding to the 3D Gaussian scene from various first viewpoints; determining the parameter sensitivity of each first Gaussian unit based on the degree of difference between the two-dimensional original images and the first rendered images corresponding to each first viewpoint; selecting second Gaussian units from all first Gaussian units whose corresponding parameter sensitivity is greater than a preset parameter sensitivity threshold; and embedding copyright information into the second Gaussian units in the 3D Gaussian scene to complete the panoramic watermark embedding operation. This application estimates the parameter sensitivity of the first Gaussian units using two-dimensional original images and first rendered images from different viewpoints to consider the geometric characteristics of the 3D scene when embedding the watermark, thus avoiding watermark information distortion. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 The flowchart illustrates a panoramic watermarking method for a three-dimensional Gaussian scene provided in an embodiment of this application. Figure 2 This document illustrates a flowchart of a watermark extraction and verification method provided in an embodiment of this application. Figure 3 This paper shows a schematic diagram of the structure of a panoramic watermarking device for a three-dimensional Gaussian scene provided in an embodiment of this application; Figure 4 A schematic diagram of the structure of an electronic device provided in an embodiment of this application is shown. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the accompanying drawings in this application are for illustrative and descriptive purposes only and are not intended to limit the scope of protection of this application. Furthermore, it should be understood that the schematic drawings are not drawn to scale. The flowcharts used in this application illustrate operations implemented according to some embodiments of this application. It should be understood that the operations in the flowcharts may not be implemented in sequence, and steps without logical contextual relationships may be reversed or implemented simultaneously. In addition, those skilled in the art, guided by the content of this application, may add one or more other operations to the flowcharts, or remove one or more operations from the flowcharts.

[0019] Furthermore, the described embodiments are merely some, not all, of the embodiments of this application. The components of the embodiments of this application described and illustrated herein can typically be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0020] To enable those skilled in the art to utilize the content of this application, and in conjunction with the specific application scenario of "3D Gaussian scene watermarking technology," the following implementation methods are provided. For those skilled in the art, the general principles defined herein can be applied to other embodiments and application scenarios without departing from the spirit and scope of this application. Although this application primarily describes the field of "3D Gaussian scene watermarking technology," it should be understood that this is merely an exemplary embodiment.

[0021] It should be noted that the term "comprising" will be used in the embodiments of this application to indicate the presence of the features declared thereafter, but does not exclude the addition of other features.

[0022] The following is a detailed description of a watermark processing method provided in the embodiments of this application.

[0023] Reference Figure 1 The diagram shown is a flowchart illustrating a panoramic watermarking method for a 3D Gaussian scene according to an embodiment of this application. The exemplary steps of this embodiment are described below: S101. Obtain the three-dimensional Gaussian scene to be processed, the two-dimensional original images of the three-dimensional Gaussian scene in each first-view perspective, and the first rendered images of the three-dimensional Gaussian scene in each first-view perspective.

[0024] In this embodiment, images of the target scene are acquired from various original viewpoints (ensuring full viewpoint coverage (surround scene or multi-angle shooting)) to obtain two-dimensional original images. Using 3D Gaussian sputtering technology, a three-dimensional visualization scene is constructed based on the two-dimensional original images from each viewpoint, resulting in the three-dimensional Gaussian scene to be processed (3D Gaussian scene). Randomly selected from all original viewpoints. A representative perspective is used to obtain the first-person perspective, denoted as... For each first-person perspective The first rendered image is generated by Gaussian splash rendering. ; where the color of each pixel in the first rendered image It is obtained by weighted superposition of the colors of Gaussian units after depth sorting; that is, each first rendered image is a two-dimensional image obtained by rendering the first Gaussian unit in the corresponding first view in the three-dimensional Gaussian scene.

[0025] Among them, the viewpoint refers to the camera shooting angle corresponding to the original multi-view image used to reconstruct the 3D Gaussian scene, which is one of the core components of the input data during the 3DGS training process. This viewpoint contains two parts of information: (1) camera pose: that is, the position (translation) and orientation (rotation) of the camera in the world coordinate system when the image is captured; (2) camera parameters: such as focal length, distortion coefficient, etc. (which determine the projection relationship of the image).

[0026] Here, a 3D Gaussian scene refers to a scene representation model composed of multiple independent three-dimensional Gaussian units (3D Gaussian elements). Each 3D Gaussian element is described by a complete set of parameters that define its geometric and optical properties. .in, Represents the 3D center coordinates of the Gaussian element, defining its specific position in the world coordinate system; It is an orthogonal rotation matrix used to control the orientation of the Gaussian element and ensure that the rotation operation does not change the size of the Gaussian element; The diagonal scaling matrix controls the scale of the Gaussian element along the x, y, and z axes, respectively. It is a set of spherical harmonic function (SH) coefficients, usually a 3rd order SH function, containing 16 coefficients, used to accurately characterize the color response of Gaussian units under different lighting conditions, ensuring the color fidelity of the rendered image; The transparency parameter controls the stacking weight of Gaussian units during the rendering process; the lower the transparency, the greater the contribution of the Gaussian unit to the final rendered color. The 3D Gaussian covariance matrix is ​​a key indicator describing the spatial distribution characteristics of 3D Gaussian units. The covariance matrix of each 3D Gaussian unit is calculated using a rotation matrix and a scaling matrix. This matrix is ​​positive definite, ensuring the rationality of the Gaussian distribution. Its eigenvalues ​​directly reflect the distribution range of Gaussian units in different directions, providing an important basis for Gaussian densification.

[0027] In addition, Gaussian Splatting Rendering is the core real-time rendering technology for 3D Gaussian scenes (3DGS), which is specifically used to quickly convert 3D Gaussian units into 2D images. It is a key method to achieve "millions of Gaussian units + photorealistic details + real-time frame rate". Core principle: Gaussian Splatting Rendering achieves efficient rendering through rasterization acceleration + Gaussian unit projection superposition: (1) Gaussian unit projection: Project each 3D Gaussian unit (an ellipsoid with position, shape, color, and transparency) onto the 2D pixel plane of the current camera and calculate its coverage area on the pixel; (2) Depth sorting: Sort the projected Gaussian units by depth (distance from the camera) to ensure that the closer Gaussians can correctly occlude the farther Gaussians; (3) Color superposition: Superimpose the color (combined with transparency) of each Gaussian unit onto the corresponding pixel in the sorted order to finally generate a 2D image.

[0028] S102. Determine the parameter sensitivity of each first Gaussian unit based on the difference between the original two-dimensional image and the first rendered image under each first viewpoint corresponding to each first Gaussian unit.

[0029] In this embodiment, the first Gaussian unit refers to a Gaussian unit in a 3D Gaussian scene viewed from a first perspective. Parameter sensitivity is used to quantify the impact of parameter perturbation introduced by embedding a watermark at the corresponding first Gaussian unit on the geometric integrity and rendering visual consistency of the 3D Gaussian scene. Specifically, the parameter sensitivity of any first Gaussian unit is calculated according to the following steps: Step 1: For each first viewpoint where the first Gaussian unit is located, calculate the parameter sensitivity matrix of the first Gaussian unit in the first viewpoint based on the degree of difference between the two-dimensional original image and the first rendered image in the first viewpoint.

[0030] In this embodiment, sensitivity quantification is achieved through approximate calculation of the Hessian matrix (i.e., the parameter sensitivity matrix). The Hessian matrix, as a second-order derivative matrix, can more accurately reflect parameter sensitivity; however, directly calculating the second derivative is extremely complex (especially when the number of Gaussian units reaches millions). Approximating the Hessian matrix using the gradient outer product method can significantly improve computational efficiency while maintaining accuracy. Therefore, the parameter sensitivity matrix is... Defined as a loss function with respect to parameters The second-order partial derivative matrix. Therefore, by substituting the difference between the original two-dimensional image and the first rendered image from the first viewpoint into the following formula, the parameter sensitivity matrix of the first Gaussian unit under the first viewpoint is obtained; ; in, The parameter sensitivity matrix of the first Gaussian element under the first viewpoint; This is the k-th first-person perspective; Let be the parameter set of the i-th first Gaussian unit; For first-person perspective Below, the difference value between the original 2D image and the first rendered image; Values ​​representing the degree of difference The parameter set of the i-th first Gaussian unit The first-order gradient vector; This is a transpose.

[0031] Here, in the construction and training of the 3D Gaussian scene, the loss function affects the parameters. The loss function in the first-order gradient is primarily photometric loss, and in some scenarios, regularization loss is added to assist in optimization. Photometric loss This is the most important loss function during training, used to measure the pixel-level difference between the "image rendered by the 3D Gaussian model" and the "real input multi-view image". The above approximation method has high accuracy after the 3DGS parameter optimization converges, because at this time the parameters are close to the optimal value, and the second derivative of the loss function can be effectively approximated by the outer product of the first derivative. Therefore, the 3D Gaussian scene obtained in this embodiment has completed parameter optimization and convergence, that is, the rendering quality of the scene has reached the practical requirements (e.g., PSNR≥25dB, SSIM≥0.8), and the number of Gaussian units and parameter distribution of the scene have stabilized, so there is no need for large-scale parameter adjustment.

[0032] Among them, PSNR (Peak Signal-to-Noise Ratio) is an indicator that measures the degree of image distortion, measured in decibels (dB). A higher value indicates less image distortion and better visual quality. SSIM (Structural Similarity Index) is an indicator that measures the structural similarity between two images, with a value ranging from [0,1]. The closer the value is to 1, the more consistent the structure, brightness, and contrast of the two images are, and the smaller the visual difference.

[0033] Step 2: Fuse the parameter sensitivity matrices of the first Gaussian element under all corresponding first-viewpoints to obtain the parameter sensitivity of the first Gaussian element.

[0034] In this embodiment, after obtaining the parameter sensitivity matrix (Hessian matrix) of the first Gaussian unit under all corresponding first viewpoints, it is necessary to further quantify the global sensitivity of each first Gaussian unit. Specifically, the parameter sensitivity matrix of the first Gaussian unit under all corresponding first viewpoints is substituted into the following formula to obtain the parameter sensitivity of the first Gaussian unit; ; in, For the parameter sensitivity of the i-th first Gaussian unit, Let i be the number of first-view units containing the i-th first Gaussian unit. The trace of the matrix (i.e., the sum of the diagonal elements of the matrix) is used to convert Hessian information in matrix form into scalar values, which facilitates subsequent threshold filtering and comparison. The parameter sensitivity matrix of the first Gaussian element under the first viewpoint. For the k-th first-person perspective, Let be the parameter set of the i-th first Gaussian unit.

[0035] Here, for each first Gaussian unit In all first-person perspectives The global parameter sensitivity of the cell is obtained by accumulating the Hessian matrix traces. The physical significance of the matrix trace lies in its ability to comprehensively reflect the overall size of the Hessian matrix, i.e., the parameters. The degree of second-order influence on the loss function; the larger the trace value, the stronger the parameter. The higher the second-order sensitivity, the stronger the tolerance to disturbances.

[0036] S103. From all the first Gaussian units, select the second Gaussian unit whose corresponding parameter sensitivity is greater than the preset parameter sensitivity threshold.

[0037] In this embodiment of the application, sensitivity to all first Gaussian element parameters is considered. The sorting process typically selects the first Gaussian unit, which has the highest parameter sensitivity (5% to 10%), as the preset parameter sensitivity threshold for filtering high-tolerance units. Filter out those that meet the requirements The first Gaussian element is used as the second Gaussian element. These Gaussian units are mainly distributed in the background areas, transition areas between objects, and smooth surface areas of a 3D Gaussian scene. The Gaussian units in these areas have a significantly higher tolerance for parameter perturbations than in other areas. Even if their parameters are moderately modified or additional denser units are added, there will be no perceptible impact on the geometry and rendering quality of the 3D Gaussian scene, making them ideal areas for watermark embedding.

[0038] Here, parameter sensitivity is the core indicator characterizing the sensitivity of 3D Gaussian parameters to the rendering result, denoted as . . The magnitude of this value directly reflects the parameter's tolerance to external disturbances: A higher value indicates a larger posterior variance for the parameter during scene optimization, a stronger tolerance to external disturbances, and a smaller impact on rendering quality after watermark embedding; conversely, a lower value indicates a lower posterior variance. The lower the value, the more sensitive the parameter is to perturbations; even minor modifications can lead to significant distortion in the rendered image. Quantifying parameter sensitivity is the key foundation for achieving watermark stealth embedding in this invention. Its calculation method is based on the Laplacian approximation in Bayesian learning and the sensitivity estimation approach of the Hessian matrix.

[0039] S104. Embed the copyright message in the second Gaussian unit in the 3D Gaussian scene to complete the panoramic watermark embedding operation.

[0040] In this embodiment of the application, the copyright message M is encoded into the distribution characteristics of the dense Gaussian unit by generating a dense Gaussian unit. Then, the dense Gaussian unit is fused with the second Gaussian unit in the three-dimensional Gaussian scene to obtain the target Gaussian scene after embedding the panoramic watermark.

[0041] Specifically, the ownership identification information of the 3D Gaussian scene to be embedded is encoded using a binary encoding format to obtain the copyright message, denoted as M. The length of the copyright message is... The maximum supported length is 32 bits. Copyright messages typically contain core information such as the asset owner's identifier, asset creation timestamp, and authorized usage permissions, ensuring integrity during message transmission.

[0042] Furthermore, based on the aforementioned watermark embedding mechanism, this application embodiment also provides a watermark extraction and verification mechanism. (Refer to...) Figure 2The diagram shown is a flowchart of a watermark extraction and verification method provided in an embodiment of this application. Specifically, after embedding the copyright message within the second Gaussian unit in a 3D Gaussian scene, the method further includes: S201. Input the parameters of each third Gaussian unit in the target 3D Gaussian scene into the 3D watermark extraction model to obtain the first copyright extraction message; the target 3D Gaussian scene is the 3D Gaussian scene after embedding the copyright message.

[0043] In this embodiment, the 3D watermark extraction model is directly derived from the target 3D Gaussian scene with the watermark. Extract copyright information. Due to... The number of Gaussian elements is typically large (hundreds of thousands to millions). Directly processing all Gaussian elements would result in extremely low extraction efficiency. Therefore, a hierarchical sampling strategy is required to randomly sample 10k to 50k Gaussian elements from the target 3D Gaussian scene to obtain the third Gaussian element. The parameters of the third Gaussian element are then defined.

[0044] Here, the network structure design of the 3D watermark extraction model adopts a convolutional neural network architecture to process the 3D Gaussian unit parameters, thus constructing the 3D watermark extraction model. This architecture is specifically designed for discrete 3D data such as point clouds, effectively extracting global features and is highly compatible with the parametric features of 3D Gaussian units. The 3D watermark extraction model consists of three stages: a feature extraction module, a feature aggregation module, and a first copyright message decoding module.

[0045] In addition, to improve the robustness of copyright message extraction, 3D anti-interference training needs to be added before inputting noise data into the 3D watermark extraction model. This simulates common 3D asset tampering attacks, ensuring that the 3D watermark extraction model can still accurately extract copyright messages even under attack conditions. The 3D anti-interference training mainly simulates four types of attacks: Gaussian noise attack, adding Gaussian noise with a standard deviation σ=0.1 to the center position of the first sampled Gaussian unit to simulate noise interference during parameter transmission; rotation attack, applying random rotation around an arbitrary axis with an angle range of ±π / 6 to all first sampled Gaussian units to simulate the scenario where the 3D asset is rotated and tampered with; translation attack, adding random translation within the range [0,1000] to the center position of all first sampled Gaussian units to simulate the scenario where the 3D asset is moved as a whole; and clipping attack, randomly removing 10% of the first sampled Gaussian units (clipping rate cr=0.1) to simulate the scenario where the 3D asset is partially deleted. The output of the anti-interference layer serves as the input to the 3D watermark extraction model. By incorporating these attack scenarios during training, the 3D watermark extraction model learns features with anti-interference capabilities, ensuring accurate extraction of copyright information even if the 3D asset is tampered with in practical applications. In other words, the aforementioned 3D watermark extraction model is pre-trained using 3D Gaussian scene samples embedded with copyright information and their corresponding real copyright information. Specifically, the first sampled Gaussian unit in the 3D Gaussian scene samples, used to input the corresponding parameters into the 3D watermark extraction model, is a Gaussian unit attacked by at least one of the following attack types: Gaussian noise attack, rotation attack, translation attack, and cropping attack.

[0046] Specifically, the parameters of each third Gaussian unit in the target 3D Gaussian scene are input into the 3D watermark extraction model according to the following steps to obtain the first copyright extraction message: Step 1: Input the parameters of each third Gaussian unit into the feature extraction module of the 3D watermark extraction model to obtain the local feature vector corresponding to each third Gaussian unit.

[0047] In this embodiment, the local eigenvector corresponding to each third Gaussian unit is used to characterize the correlation between parameters within the third Gaussian unit, and / or the correlation between the parameters of the third Gaussian unit and the parameters of its neighboring third Gaussian units. For example, the orthogonal rotation matrix within the third Gaussian unit... and diagonal scaling matrix The correlation, whether the colors of adjacent third Gaussian units are continuous, etc.

[0048] Here, the feature extraction module consists of three fully connected layers. The parameters of each third Gaussian unit pass through these three fully connected layers sequentially. The feature vectors output by each fully connected layer have dimensions of 64, 128, and 128 respectively to extract local parameter correlations. The output of the last fully connected layer is determined as the final local feature vector.

[0049] Step 2: Input the local feature vectors corresponding to all third Gaussian units into the feature aggregation module in the 3D watermark extraction model to obtain the first global feature vector corresponding to the target 3D Gaussian scene.

[0050] In this embodiment of the application, the local feature vectors corresponding to all third Gaussian units are aggregated by max pooling to obtain a 128-dimensional first global feature vector.

[0051] Step 3: Input the first global feature vector into the first watermark decoding module in the 3D watermark extraction model to obtain the first copyright extraction message.

[0052] In this embodiment, the first global feature vector is input into the first fully connected layer, and a 96-dimensional first feature vector is output. The 96-dimensional first feature vector is then input into the second fully connected layer and mapped to a first binary prediction vector with the same length as the embedded copyright message. Finally, the first binary prediction vector is mapped to the [0,1] interval through the Sigmoid activation function and output. The output value is predicted as 1 if it is greater than a preset threshold (0.5), and 0 otherwise, thus obtaining the binary first copyright extraction message.

[0053] S202, and / or, input all the second rendered images corresponding to the target 3D Gaussian scene into the 2D watermark extraction model to obtain the second copyright extraction message; each second rendered image is a 2D image obtained by rendering the fourth Gaussian unit in the target 3D Gaussian scene under the corresponding second viewpoint.

[0054] In this embodiment of the application, the two-dimensional watermark extraction model extracts watermarks from the target three-dimensional Gaussian scene. The second set of rendered images Extracting copyright information requires careful consideration of the inherent distortion characteristics of 2D images. Reliable extraction is achieved through image preprocessing, a dedicated network structure, anti-interference layer design, and multi-view message fusion. The second rendered image set refers to a set of 2D rendered images generated using Gaussian splashing technology on a target 3D Gaussian scene. These images are generated by uniformly selecting multiple panoramic viewpoints (covering a 360° horizontal view and a 90° vertical view) within a spherical coordinate system with the scene center as the origin. The second rendered image set contains second rendered images. The resolution of each second rendered image can be set according to actual application requirements, typically ranging from 1024×768 to 2048×1536, ensuring complete coverage of all key areas of the scene. The fourth Gaussian unit refers to the Gaussian unit within the target 3D Gaussian scene.

[0055] In addition, to improve the robustness of the 2D watermark extraction model, 2D anti-interference data is added during image preprocessing to simulate common 2D image tampering attacks, mainly including four types of attacks: JPEG compression attacks, using quality factors. The model employs various attack scenarios: JPEG compression (JPEG compression is the most common distortion method in image transmission; a quality factor of 50 corresponds to a medium level of compression, which introduces some blockiness); scaling attacks (scaling the image to 25% of its original size and then enlarging it back to its original size using bilinear interpolation) to simulate a scaled and tampered image; Gaussian blur attacks (using a 3×3 Gaussian blur kernel with a standard deviation of 0.1 to blur the image, simulating blur distortion during image transmission); and noise attacks (adding Gaussian noise with a standard deviation of 0.05 to simulate image sensor noise or transmission noise). By incorporating these attack scenarios into the training process, the decoder learns image features with anti-interference capabilities, ensuring accurate extraction of copyright information even if the 2D rendered image is tampered with. In other words, the above 2D watermark extraction model is pre-trained using rendered sample images corresponding to 3D Gaussian scene samples embedded with copyright information (obtained in the same way as the second rendered image) and real copyright information. The rendered sample images are images attacked by at least one of the following attack types: JPEG compression attack, scaling attack, Gaussian blur attack, and noise attack.

[0056] Specifically, the second rendered images corresponding to the target 3D Gaussian scene are input into the 2D watermark extraction model according to the following steps to obtain the second copyright extraction message, including: Step 1: Input each of the second rendered images into the multi-scale feature extraction layer of the two-dimensional watermark extraction model to obtain the first image feature map of each second rendered image at each preset scale.

[0057] In this embodiment, the multi-scale feature extraction layer includes four convolutional blocks, each used to extract a first image feature map at a corresponding preset scale to capture multi-scale pixel information. The preset scales include 512×384×64, 256×192×128, 128×96×256, and 32×24×512.

[0058] Step 2: For each second rendered image, input the first image feature map of the second rendered image at all preset scales into the first feature fusion layer of the two-dimensional watermark extraction model to obtain the second image feature vector corresponding to the second rendered image.

[0059] In this embodiment, the first image feature map of the second rendered image at all preset scales is average pooled and compressed into a 512-dimensional single-view rendering image of global features, resulting in a second image feature vector. This second image feature vector can comprehensively reflect the overall features of the second rendered image, including subtle feature differences formed by the watermark Gaussian in the rendered image.

[0060] Step 3: Input the second image feature maps corresponding to all the second rendered images into the second feature fusion layer of the two-dimensional watermark extraction model to obtain the second global feature vector corresponding to the target three-dimensional Gaussian scene.

[0061] Step 4: Input the second global feature vector into the second copyright message decoding module in the two-dimensional watermark extraction model to obtain the second copyright extraction message.

[0062] In this embodiment, the second global feature vector is input into the first fully connected layer, which outputs a 96-dimensional second feature vector. The 96-dimensional second feature vector is then input into the second fully connected layer and mapped to a second binary prediction vector with a length consistent with the embedded copyright message. Finally, the second binary prediction vector is mapped to the [0,1] interval using the Sigmoid activation function and output. Output values ​​greater than a preset threshold (0.5) are predicted as 1, and otherwise as 0, thus obtaining the binary second copyright extraction message.

[0063] Here, the core objective of the watermark extraction model is to reliably extract copyright information from a watermarked 3D Gaussian scene and its second rendered image set, supporting both 3D and 2D dual-end verification to address different application scenarios. The 3D watermark extraction model is suitable for scenarios where 3D assets are not rendered and only Gaussian parameters are retained (such as 3D asset transactions and cloud storage verification), while the 2D watermark extraction model is suitable for scenarios where 3D assets have been rendered as 2D images and 3D parameters cannot be obtained (such as social media dissemination and offline display verification). By designing dedicated 3D and 2D watermark extraction models, combined with an anti-interference layer design, it is ensured that copyright information can still be accurately extracted even in the event of tampering attacks.

[0064] S203. Based on the matching degree between the first copyright extraction message and the embedded copyright message, and / or the matching degree between the second copyright extraction message and the embedded copyright message, determine the watermark verification result corresponding to the target 3D Gaussian scene.

[0065] In this application embodiment, if the matching degree is greater than the preset matching threshold (such as 95%), the watermark verification result corresponding to the target 3D Gaussian scene is verified as passed; otherwise, the verification fails.

[0066] When determining the watermark verification result corresponding to the target 3D Gaussian scene based on the matching degree between the first copyright extraction message and the embedded copyright message, and the matching degree between the second copyright extraction message and the embedded copyright message, the two matching degrees can be weighted and fused or averaged to determine the target matching degree, which is then compared with a preset matching threshold to obtain the watermark verification result corresponding to the target 3D Gaussian scene. Alternatively, if either matching degree is greater than the preset matching threshold, the watermark verification result corresponding to the target 3D Gaussian scene is considered to have passed verification; otherwise, it is considered to have failed verification.

[0067] Here, the expression for calculating the matching degree is: ,in This is an indicator function that takes the value 1 when the condition inside the parentheses is true, and 0 otherwise. For first copyright extraction message or second copyright extraction message The value of the t-th bit in the middle. For embedded copyright messages The value of the t-th bit. For embedded copyright messages The length.

[0068] Based on the same inventive concept, this application also provides a panoramic watermarking device for a three-dimensional Gaussian scene corresponding to the panoramic watermarking method for a three-dimensional Gaussian scene. Since the principle of the device in this application is similar to the panoramic watermarking method for a three-dimensional Gaussian scene described above in this application, the implementation of the device can refer to the implementation of the method, and the repeated parts will not be described again.

[0069] Reference Figure 3 The diagram shown is a schematic of a panoramic watermarking device for a three-dimensional Gaussian scene provided in an embodiment of this application. The device includes: The acquisition module 301 is used to acquire the three-dimensional Gaussian scene to be processed, to construct the two-dimensional original images of the three-dimensional Gaussian scene in each first view, and the first rendered images of the three-dimensional Gaussian scene in each first view; each first rendered image is a two-dimensional image obtained by image rendering based on the first Gaussian unit in the three-dimensional Gaussian scene in the corresponding first view. The sensitivity determination module 302 is used to determine the parameter sensitivity of each first Gaussian unit based on the degree of difference between the two-dimensional original image and the first rendered image under each first viewpoint corresponding to each first Gaussian unit; the parameter sensitivity is used to quantify the degree of influence of the parameter perturbation introduced by embedding watermark at the corresponding first Gaussian unit on the geometric integrity and rendering visual consistency of the three-dimensional Gaussian scene. The Gaussian unit filtering module 303 is used to filter second Gaussian units from all first Gaussian units whose corresponding parameter sensitivity is greater than a preset parameter sensitivity threshold. The watermark embedding module 304 is used to embed copyright information in the second Gaussian unit in the three-dimensional Gaussian scene, thereby completing the panoramic watermark embedding operation.

[0070] In one possible implementation, the sensitivity determination module 302 is specifically configured to calculate the parameter sensitivity of any first Gaussian unit according to the following steps: for each first viewpoint where the first Gaussian unit is located, calculate the parameter sensitivity matrix of the first Gaussian unit in the first viewpoint according to the difference value between the two-dimensional original image and the first rendered image in the first viewpoint; fuse the parameter sensitivity matrices of the first Gaussian unit in all corresponding first viewpoints to obtain the parameter sensitivity of the first Gaussian unit.

[0071] In one possible implementation, the sensitivity determination module 302 is specifically used to substitute the difference value between the two-dimensional original image and the first rendered image under the first viewpoint into the following formula to obtain the parameter sensitivity matrix of the first Gaussian unit under the first viewpoint. ; in, The parameter sensitivity matrix of the first Gaussian element under the first viewpoint; This is the k-th first-person perspective; Let be the parameter set of the i-th first Gaussian unit; For first-person perspective Below, the difference value between the original 2D image and the first rendered image; Values ​​representing the degree of difference The parameter set of the i-th first Gaussian unit The first-order gradient vector; This is a transpose.

[0072] In one possible implementation, the sensitivity determination module 302 is specifically used to substitute the parameter sensitivity matrix of the first Gaussian unit under all corresponding first viewpoints into the following formula to obtain the parameter sensitivity of the first Gaussian unit. ; in, For the parameter sensitivity of the i-th first Gaussian unit, Let i be the number of first-view units containing the i-th first Gaussian unit. The trace of the matrix, The parameter sensitivity matrix of the first Gaussian element under the first viewpoint. For the k-th first-person perspective, Let be the parameter set of the i-th first Gaussian unit.

[0073] In one possible implementation, the device further includes a watermark extraction and verification module 305; after embedding the copyright message in the second Gaussian unit in the 3D Gaussian scene, the watermark extraction and verification module 305 is used to input the parameters of each third Gaussian unit in the target 3D Gaussian scene into a 3D watermark extraction model to obtain a first copyright extraction message; the target 3D Gaussian scene is the 3D Gaussian scene after embedding the copyright message; and / or, input all the second rendered images corresponding to the target 3D Gaussian scene into a 2D watermark extraction model to obtain a second copyright extraction message; each second rendered image is a 2D image obtained by rendering the fourth Gaussian unit in the target 3D Gaussian scene under the corresponding second viewpoint; the watermark verification result corresponding to the target 3D Gaussian scene is determined according to the matching degree between the first copyright extraction message and the embedded copyright message, and / or the matching degree between the second copyright extraction message and the embedded copyright message.

[0074] In one possible implementation, the watermark extraction and verification module 305 is specifically used to input the parameters of each third Gaussian unit into the feature extraction module in the three-dimensional watermark extraction model to obtain the local feature vector corresponding to each third Gaussian unit; the local feature vector corresponding to each third Gaussian unit is used to characterize the correlation between the parameters within the third Gaussian unit, and / or the correlation between the parameters of the third Gaussian unit and the parameters of its neighboring third Gaussian units; input the local feature vectors corresponding to all third Gaussian units into the feature aggregation module in the three-dimensional watermark extraction model to obtain the first global feature vector corresponding to the target three-dimensional Gaussian scene; input the first global feature vector into the first copyright message decoding module in the three-dimensional watermark extraction model to obtain the first copyright extraction message.

[0075] In one possible implementation, the watermark extraction and verification module 305 is specifically used to input each second rendered image into the multi-scale feature extraction layer of the two-dimensional watermark extraction model to obtain a first image feature map of each second rendered image at each preset scale; for each second rendered image, input the first image feature map of the second rendered image at all preset scales into the first feature fusion layer of the two-dimensional watermark extraction model to obtain a second image feature vector corresponding to the second rendered image; input the second image feature vectors corresponding to all second rendered images into the second feature fusion layer of the two-dimensional watermark extraction model to obtain a second global feature vector corresponding to the target three-dimensional Gaussian scene; and input the second global feature vector into the second copyright message decoding module of the two-dimensional watermark extraction model to obtain a second copyright extraction message.

[0076] Here, this application embodiment provides a panoramic watermarking device for a three-dimensional Gaussian scene. This device estimates the parameter sensitivity of the first Gaussian unit by using two-dimensional original images and a first rendered image from different perspectives, so as to take into account the geometric structural characteristics of the three-dimensional scene when embedding the watermark and avoid distortion of the watermark information.

[0077] like Figure 4 As shown in the embodiment of this application, an electronic device 400 includes a processor 401, a memory 402, and a bus. The memory 402 stores machine-readable instructions that can be executed by the processor 401. When the electronic device is running, the processor 401 communicates with the memory 402 via the bus. The processor 401 executes the machine-readable instructions to perform the steps of the panoramic watermarking method for the three-dimensional Gaussian scene described above.

[0078] Specifically, the memory 402 and processor 401 mentioned above can be general-purpose memory and processor, without any specific limitations. When the processor 401 runs the computer program stored in the memory 402, it can execute the above-mentioned panoramic watermarking processing method for three-dimensional Gaussian scenes.

[0079] Corresponding to the above-described panoramic watermarking method for three-dimensional Gaussian scenes, this application embodiment also provides a computer-readable storage medium storing a computer program, which, when run by a processor, executes the steps of the above-described panoramic watermarking method for three-dimensional Gaussian scenes.

[0080] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and devices described above can be referred to the corresponding processes in the method embodiments, and will not be repeated here. In the several embodiments provided in this application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed mutual coupling or direct coupling or communication connection can be through some communication interfaces; the indirect coupling or communication connection of devices or modules can be electrical, mechanical, or other forms.

[0081] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0082] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0083] If the aforementioned function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a processor-executable, non-volatile, computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the panoramic watermarking method for the three-dimensional Gaussian scene described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, external hard drives, ROM, RAM, magnetic disks, or optical disks.

[0084] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A panorama watermarking method of a three-dimensional Gaussian scene, characterized by, The method comprises: acquiring a three-dimensional Gaussian scene to be processed, a two-dimensional original image corresponding to each first view angle of the three-dimensional Gaussian scene, and a first rendered image corresponding to each first view angle of the three-dimensional Gaussian scene; each first rendered image is a two-dimensional image obtained by image rendering of a first Gaussian unit in the three-dimensional Gaussian scene at a corresponding first view angle; determining a parameter sensitivity of each first Gaussian unit according to a difference degree value between the two-dimensional original image and the first rendered image at each first view angle of each first Gaussian unit; the parameter sensitivity is used to quantify the influence degree of parameter disturbance introduced by embedding a watermark at a corresponding first Gaussian unit on the geometric structure integrity and the rendering visual consistency of the three-dimensional Gaussian scene; from all the first Gaussian units, filtering a second Gaussian unit corresponding to a parameter sensitivity greater than a preset parameter sensitivity threshold; embedding a copyright message in the second Gaussian unit in the three-dimensional Gaussian scene, thereby completing the embedding operation of the panoramic watermark.

2. The method of claim 1, wherein, The parameter sensitivity of any first Gaussian unit is calculated according to the following steps, comprising: for each first view angle in which the first Gaussian unit is located, calculating a parameter sensitivity matrix of the first Gaussian unit at the first view angle according to a difference degree value between the two-dimensional original image and the first rendered image at the first view angle; fusing the parameter sensitivity matrices of the first Gaussian unit at all corresponding first view angles to obtain the parameter sensitivity of the first Gaussian unit.

3. The method of claim 2, wherein, The parameter sensitivity matrix of the first Gaussian unit at the first view angle is calculated according to a difference degree value between the two-dimensional original image and the first rendered image at the first view angle, comprising: substituting the difference degree value between the two-dimensional original image and the first rendered image at the first view angle into the following formula to obtain the parameter sensitivity matrix of the first Gaussian unit at the first view angle; ； in, The parameter sensitivity matrix of the first Gaussian element under the first viewpoint; This is the k-th first-person perspective; Let be the parameter set of the i-th first Gaussian unit; For first-person perspective Below, the difference value between the original 2D image and the first rendered image; Values ​​representing the degree of difference The parameter set of the i-th first Gaussian unit The first-order gradient vector; This is a transpose.

4. The method of claim 2, wherein, The parameter sensitivity of the first Gaussian unit is fused by substituting the parameter sensitivity matrices of the first Gaussian unit at all corresponding first view angles into the following formula to obtain the parameter sensitivity of the first Gaussian unit. After the copyright message is embedded in the second Gaussian unit in the three-dimensional Gaussian scene, the method further comprises: ； wherein, is a parameter sensitivity of the i-th first Gaussian element, is a number of first views in which the i-th first Gaussian element is located, is a trace of a matrix, is a parameter sensitivity matrix of the first Gaussian element in the first view, is the k-th first view, is a parameter set of the i-th first Gaussian element.

5. The method of claim 1, wherein, inputting each third Gaussian unit parameter in a target three-dimensional Gaussian scene into a three-dimensional watermark extraction model to obtain a first copyright extraction message; the target three-dimensional Gaussian scene is the three-dimensional Gaussian scene after the copyright message is embedded; and / or, inputting all second rendered images corresponding to the target three-dimensional Gaussian scene into a two-dimensional watermark extraction model to obtain a second copyright extraction message; each second rendered image is a two-dimensional image obtained by image rendering of a fourth Gaussian unit in the target three-dimensional Gaussian scene at a corresponding second view angle; ​ According to the matching degree between the first copyright removal message and the embedded copyright message, and / or the matching degree between the second copyright removal message and the embedded copyright message, a watermark verification result corresponding to the target three-dimensional Gaussian scene is determined.

6. The method of claim 5, wherein, The inputting of each third Gaussian unit parameter in the target three-dimensional Gaussian scene into the three-dimensional watermark extraction model to obtain a first copyright removal message comprises: The inputting of each third Gaussian unit parameter into a feature extraction module in the three-dimensional watermark extraction model to obtain a local feature vector corresponding to each third Gaussian unit; the local feature vector corresponding to each third Gaussian unit is used to represent the correlation between parameters inside the third Gaussian unit, and / or the correlation between the parameters of the third Gaussian unit and the parameters of the adjacent third Gaussian unit; The inputting of the local feature vectors corresponding to all third Gaussian units into a feature aggregation module in the three-dimensional watermark extraction model to obtain a first global feature vector corresponding to the target three-dimensional Gaussian scene; The inputting of the first global feature vector into a first copyright message decoding module in the three-dimensional watermark extraction model to obtain a first copyright removal message.

7. The method of claim 5, wherein the method further comprises: The inputting of all second rendered images corresponding to the target three-dimensional Gaussian scene into a two-dimensional watermark extraction model to obtain a second copyright removal message comprises: The inputting of each second rendered image into a multi-scale feature extraction layer in the two-dimensional watermark extraction model to obtain a first image feature map of each second rendered image at each preset scale; For each second rendered image, the inputting of the first image feature maps of the second rendered image at all preset scales into a first feature fusion layer in the two-dimensional watermark extraction model to obtain a second image feature vector corresponding to the second rendered image; The inputting of the second image feature vectors corresponding to all second rendered images into a second feature fusion layer in the two-dimensional watermark extraction model to obtain a second global feature vector corresponding to the target three-dimensional Gaussian scene; The inputting of the second global feature vector into a second copyright message decoding module in the two-dimensional watermark extraction model to obtain a second copyright removal message.

8. An apparatus for panorama watermarking of a three-dimensional Gaussian scene, characterized by The device comprises: An acquisition module is configured to acquire a three-dimensional Gaussian scene to be processed, two-dimensional original images corresponding to the three-dimensional Gaussian scene at each first view angle, and first rendered images corresponding to the three-dimensional Gaussian scene at each first view angle; each first rendered image is a two-dimensional image obtained by image rendering on a first Gaussian unit in the three-dimensional Gaussian scene at a corresponding first view angle; A sensitivity determination module is configured to determine the parameter sensitivity of each first Gaussian unit according to the difference degree value between the two-dimensional original image and the first rendered image at each first view angle corresponding to each first Gaussian unit; the parameter sensitivity is used to quantify the influence degree of the parameter disturbance introduced by embedding a watermark at the corresponding first Gaussian unit on the geometric structure integrity and the rendering visual consistency of the three-dimensional Gaussian scene; A Gaussian unit screening module is configured to screen second Gaussian units with a parameter sensitivity greater than a preset parameter sensitivity threshold from all first Gaussian units. The watermark embedding module is configured to embed a copyright message in a second Gaussian cell in the three-dimensional Gaussian scene, thereby completing embedding of the panoramic watermark.

9. An electronic device, comprising: The method comprises the following steps: A processor, a storage medium and a bus, the storage medium stores machine readable instructions executable by the processor, when the electronic device is running, the processor and the storage medium communicate through the bus, the processor executes the machine readable instructions to execute the steps of the panoramic watermark processing method of the three-dimensional Gaussian scene as claimed in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, when the computer program is run by the processor, the steps of the panoramic watermark processing method of the three-dimensional Gaussian scene as claimed in any one of claims 1 to 7 are executed.

Images