Trusted three-dimensional digital content generation method, system, device and medium based on physical true value

Abstract

This invention discloses an artificial intelligence constraint generation method based on physical truth, comprising: acquiring three-dimensional geometric data and visual texture data of a target object in real time via a data acquisition terminal; performing local precision verification on the three-dimensional geometric data; dynamically generating guidance information if the verification fails; and digitally signing the acquired data to obtain digital signature information if the verification passes. A controlled three-dimensional data structure is generated based on the verified three-dimensional geometric data, compressed visual texture data, and digital signature information. A cloud server verifies the legality of the digital signature of the controlled three-dimensional data structure, parses the three-dimensional geometric data in the controlled three-dimensional data structure, converts the parsed three-dimensional geometric data into spatial constraints, and sends the spatial constraints to a generative artificial intelligence model. The generative artificial intelligence model generates content that conforms to physical truth and has a reliable source based on the spatial constraints. This method solves the problems of geometric illusion, verification lag, and lack of data trust.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, specifically to a method, system, device, and medium for generating reliable 3D digital content based on physical truth. Background Technology

[0002] With the widespread application of generative artificial intelligence (AIGC) technology in fields such as 3D modeling and digital content creation, the efficiency of generating 3D digital content has been greatly improved. However, it also faces three core technical pain points, and existing technologies are unable to form effective solutions.

[0003] First, the fundamental flaws in generative AI's underlying principles lead to physical truth bias. Current mainstream generative AI models (such as those based on Latent Diffusion) are essentially generated based on the statistical probability distribution of pixels / voxels. Their core flaw lies in their inability to understand the topological structure and physical causality of three-dimensional Euclidean space. When handling Relative Physical Objects (RWA) tasks, AI often fills in high-frequency features based on statistical regularities, ignoring the object's unique dimensions, proportions, and manufacturing tolerances. Specifically, this manifests as:

[0004] A. Size and proportion distortion: The lack of absolute scale reference causes the generated industrial parts or footwear products to have aspect ratio drift, and the structure does not conform to the mechanical logic;

[0005] B. Failure to decouple texture and material: AI has difficulty distinguishing the inherent texture of physical surfaces from the reflection of ambient light and shadow, often incorrectly baking highlights as the base color, causing digital models to lose their realistic physical texture in heavily lit environments;

[0006] C. Detail smoothing: The diffusion model tends to generate idealized surfaces, automatically erasing the micro-irregularities of the entity caused by wear or process, causing the digital twin to degenerate into an ideal model.

[0007] For example, in the scenario of digitizing physical entities, the digitized content generated by artificial intelligence often has size deviations and structural misalignments, which makes it impossible for the generated content to match the true value of the physical object. The geometric consistency accuracy between the content generated by artificial intelligence and the physical object is only about 60%.

[0008] Secondly, there is a lack of real-time quality inspection and a delay in verification. Traditional solutions, such as the method mentioned in the paper "Research on Cloud-based 3D Data Verification Technology" (Computer Engineering and Applications, Vol. 49, 2023), require uploading complete 3D geometric data to the cloud before accuracy verification. This not only consumes a large amount of network bandwidth but also results in verification delays and makes it impossible to intercept invalid data in real time during the data acquisition stage.

[0009] Third, there is a lack of data trust and ownership confirmation. The existing 3D digital content generation process lacks reliable traceability and signature mechanisms. For example, the data generated by mainstream 3D scanning tools (such as Artec Eva) is not bound to hardware-level signatures, making it easy to tamper with. It is difficult to form a trust chain of data integrity verification, immutability, and traceability. Moreover, the acquisition end is prone to uploading low-quality or forged data, resulting in waste of cloud computing resources and bandwidth congestion, which restricts the compliant circulation and application of 3D digital content. Summary of the Invention

[0010] The purpose of this invention is to provide a method, system, device, and medium for generating reliable 3D digital content based on physical truth, thereby solving the problems of geometric illusion, verification lag, and lack of data trust.

[0011] This invention is achieved through the following technical solution:

[0012] In a first aspect, embodiments of the present invention provide a method for generating reliable 3D digital content based on physical truth, comprising:

[0013] The acquisition terminal acquires the three-dimensional geometric data and visual texture data of the target object in real time. The visual texture data is compressed to obtain compressed visual texture data. The three-dimensional geometric data is locally verified to obtain a verification result. If the verification result is unsuccessful, guidance information is dynamically generated. If the verification result is successful, the acquired data is digitally signed to obtain digital signature information. Based on the verified three-dimensional geometric data, compressed visual texture data and digital signature information, a controlled three-dimensional data structure is generated and sent to the cloud server.

[0014] The cloud server receives and verifies the legality of the digital signature of the controlled 3D data structure, parses the 3D geometric data in the controlled 3D data structure to obtain the parsed 3D geometric data, converts the parsed 3D geometric data into spatial constraints, and sends the spatial constraints to the generative artificial intelligence model.

[0015] Generative artificial intelligence models generate content that conforms to physical truth and has a reliable source based on spatial constraints.

[0016] Furthermore, the specific method for obtaining the verification result by performing local accuracy verification on the three-dimensional geometric data includes:

[0017] A sparse sampling strategy is used to extract feature points from the three-dimensional geometric data to obtain a feature point set;

[0018] The error is compared between the feature point set and the 3D model reconstructed locally from the visual texture data collected at the same time. If the error is less than the preset truth tolerance threshold, the verification is considered successful.

[0019] Furthermore, the sparse sampling strategy employs an adaptive dynamic step size, dynamically adjusting the sampling density based on the real-time load status of the graphics processor of the acquisition terminal.

[0020] Furthermore, the specific method for digitally signing the acquired data includes:

[0021] Calculate the cryptographic hash values ​​of the 3D geometric data and the compressed visual texture data respectively to generate a data fingerprint;

[0022] Within a trusted execution environment, a pre-defined private key is used to digitally sign the data fingerprint, generating digital signature information.

[0023] Furthermore, the acquisition terminal acquires the three-dimensional geometric data of the target object in real time, including anti-counterfeiting liveness detection of the target object, specifically including:

[0024] Acquire multi-frame 3D geometric data of the target object in a continuous time series and lock the sampling points in a static coordinate system;

[0025] Calculate the time dimension variance of the sampling point depth values ​​and perform frequency domain analysis on the trend of the time dimension variance;

[0026] If a periodic peak related to the refresh rate and frequency multiplication factor of the display device is detected in the fluctuation spectrum of the depth value, or if moiré fringe interference features are present, then the target object is determined to be a secondary imaging device that is not a real physical entity, and the generation of the controlled three-dimensional data structure is terminated.

[0027] Secondly, another embodiment of the present invention provides a reliable 3D digital content generation system based on physical truth, comprising: a data acquisition terminal, a cloud server, and a generative artificial intelligence model;

[0028] The acquisition terminal is used to acquire the three-dimensional geometric data and visual texture data of the target object in real time, compress the visual texture data to obtain compressed visual texture data, perform local precision verification on the three-dimensional geometric data to obtain a verification result, if the verification result is unsuccessful, guide information is dynamically generated, if the verification result is successful, digital signature is performed on the acquired data to obtain digital signature information, and a controlled three-dimensional data structure is generated based on the verified three-dimensional geometric data, compressed visual texture data and digital signature information, and the controlled three-dimensional data structure is sent to the cloud server.

[0029] The cloud server is used to receive and verify the legality of the digital signature of the controlled three-dimensional data structure, parse the three-dimensional geometric data in the controlled three-dimensional data structure to obtain the parsed three-dimensional geometric data, convert the parsed three-dimensional geometric data into spatial constraints, and send the spatial constraints to the generative artificial intelligence model.

[0030] The generative artificial intelligence model is used to generate content that conforms to physical truth and has a reliable source, based on spatial constraints.

[0031] Furthermore, the acquisition terminal includes a local verification module. The local verification module uses a sparse sampling strategy to extract feature points from the three-dimensional geometric data to obtain a feature point set. The feature point set is then compared with the three-dimensional model reconstructed locally from the visual texture data acquired at the same time. If the error is less than a preset truth tolerance threshold, the verification is deemed successful.

[0032] Furthermore, the acquisition terminal includes a hardware security module, which is used to calculate the encrypted hash values ​​of the three-dimensional geometric data and the compressed visual texture data respectively, generate a data fingerprint, and digitally sign the data fingerprint using a preset private key in a trusted execution environment to generate digital signature information.

[0033] Thirdly, another embodiment of the present invention provides an electronic device comprising: a processor, an input device, an output device, and a memory, wherein the processor, the input device, the output device, and the memory are interconnected, the memory is used to store a computer program, the computer program includes program instructions, and the processor is configured to invoke the program instructions to execute the method described in the first embodiment above.

[0034] Fourthly, another embodiment of the present invention provides a computer-readable storage medium storing a computer program, the computer program including program instructions that, when executed by a processor, cause the processor to perform the method described in the first embodiment above.

[0035] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0036] This invention provides a method, system, device, and medium for generating trusted 3D digital content based on physical truth values. This significantly improves the accuracy of geometric consistency between AI-generated content and physical objects, multiplies processing efficiency, and adapts to various scenarios such as physical entity digitization, high-precision digital twin construction, and industrial design. By digitally signing the real-time acquired data, not only is the data source proven, but the digital signature information is also used as a key to generate a controlled 3D data structure with the 3D geometric data. This structure serves as a spatial constraint for the generative AI model. The cloud server verifies the signature validity of this constraint before generation can proceed, thus eliminating the possibility of using unverified or tampered data as constraints at the algorithmic logic level.

[0037] Hardware-level signature and hash binding mechanism, combined with end-to-end traceability design, enables trusted 3D digital content to have traceability and anti-tampering capabilities. This ensures the authenticity of the source and data integrity of trusted 3D digital content during transmission in untrusted network environments, prevents man-in-the-middle attacks or data injection, and provides judicial-grade evidentiary value to support scenarios such as copyright protection and compliant circulation of trusted 3D digital content, thus solving the data trust problem.

[0038] Because digital signatures and generative AI models are deeply bound to each other at the data structure level, the copyright of any content generated based on these constraints can be traced back to the data collection source represented by the signing hardware. The signature information includes the hardware identifier and timestamp of the data collection terminal, serving as a link in the chain of evidence and enabling copyright anchoring from data to generated content. This solves the problem of digital asset ownership verification and provides native technological protection for the compliant use of digitized physical entities. Attached Figure Description

[0039] To more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of the present invention and should not be considered 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. In the drawings:

[0040] Figure 1 A flowchart of a reliable 3D digital content generation method based on physical truth provided in the first embodiment of the present invention;

[0041] Figure 2 A structural diagram of a controlled three-dimensional data structure;

[0042] Figure 3 The present invention provides a structural block diagram of a reliable 3D digital content generation system based on physical truth values, which is another embodiment of the present invention. Detailed Implementation

[0043] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the embodiments and accompanying drawings. The illustrative embodiments and descriptions of the present invention are only used to explain the present invention and are not intended to limit the present invention.

[0044] like Figure 1 As shown, the first embodiment of the present invention provides a method for generating reliable 3D digital content based on physical truth, comprising:

[0045] The acquisition terminal acquires the three-dimensional geometric data and visual texture data of the target object in real time. The visual texture data is compressed to obtain compressed visual texture data. The three-dimensional geometric data is locally verified to obtain a verification result. If the verification result is unsuccessful, guidance information is dynamically generated. If the verification result is successful, the acquired data is digitally signed to obtain digital signature information. Based on the verified three-dimensional geometric data, compressed visual texture data, and digital signature information, a controlled three-dimensional data structure is generated and sent to the cloud server.

[0046] The cloud server receives and verifies the legality of the digital signature of the controlled 3D data structure, parses the 3D geometric data in the controlled 3D data structure to obtain the parsed 3D geometric data, converts the parsed 3D geometric data into spatial constraints, and sends the spatial constraints to the generative artificial intelligence model.

[0047] Generative artificial intelligence models generate content that conforms to physical truth and has a reliable source based on spatial constraints.

[0048] The data acquisition terminal utilizes a smartphone or tablet equipped with a LiDAR depth sensor, a neural network processing unit, and a GPU. It integrates a data acquisition module, a local verification module, a hardware security module, and a trusted 3D digital content generation module. The data acquisition module includes a LiDAR depth sensor and an image sensor, which simultaneously acquire 3D geometric and visual texture data of the target object. The local verification module extracts feature points using an adaptive sparse sampling strategy and calculates the root mean square error (RMSE) of the model reconstructed from the synchronously acquired visual data to verify accuracy. The verification result can dynamically generate AR guidance information to prompt the user to re-acquire unqualified areas. Based on the verification result, guidance information is generated, achieving an interactive closed loop of acquisition, verification, and re-acquisition, enhancing the effectiveness of invalid data interception. Local interactive re-acquisition improves data quality from the source, avoiding subsequent duplicate verification and invalid uploads. The hardware security module uses the ECDSA P-256 algorithm for signing, binding the signature information to the hash value of the controlled 3D data structure, achieving coupling with AI-generated constraints. The trusted 3D digital content generation module packages the verified 3D geometric data, compressed texture data, and signature information to form a controlled 3D data structure.

[0049] Upon receiving the user's acquisition command, the acquisition terminal wakes up the LiDAR depth sensor and image sensor. Through hardware timestamp alignment technology, it locks the time deviation between the depth frame and the visual frame within the millisecond range, ensuring the spatial consistency of geometric data and texture data.

[0050] Before formal recording, the system utilizes the light sensor and accelerometer of the data acquisition terminal to monitor ambient brightness and device stability in real time. If the ambient brightness is lower than the preset light sensitivity threshold, or if the device jitter exceeds the preset motion blur threshold, the system will pause the data acquisition logic and output a status prompt.

[0051] During the data acquisition process, the acquisition terminal defaults to enabling a micro-motion feature analysis background process for anti-spoofing liveness detection. This process analyzes the noise distribution characteristics of multiple consecutive frames of point cloud data to identify whether there are microscopic static features on the surface of the target object that conform to physical laws. If the detected point cloud noise exhibits a periodic pattern strongly correlated with the display device's refresh rate, the target is determined to be a secondary imaging device that is not a real physical entity. The generation of the controlled 3D data structure is then terminated, blocking the generation of false data at its source. The specific method of micro-motion feature analysis includes: taking the depth value sequence D(t) of sampling points P in the same static coordinate system in N consecutive frames (e.g., N=30) of point cloud. The variance of D(t) is calculated and a Fast Fourier Transform (FFT) is performed. The spectrum is analyzed; if a significant peak exceeding the background noise level β (e.g., 20dB) appears at the common display refresh rate fundamental frequency (e.g., 60Hz, 120Hz) and its second harmonic (120Hz, 240Hz), a screen imaging pattern is determined to exist. At the same time, the ratio of the sensor sampling frequency to the suspected refresh rate is calculated. If it is close to a rational number, it is determined that moiré fringe interference may occur. The non-physical determination is made by combining the two results.

[0052] Cloud server: Receives controlled 3D data structure sent by the acquisition terminal, parses the 3D geometric data and signature information in the packet, verifies the legality of the signature, converts the 3D geometric data into spatial constraints that can be recognized by artificial intelligence, and sends them to the generative artificial intelligence model.

[0053] Generative AI model: Receives spatial constraints from cloud server, generates content that conforms to physical truth based on 3D geometric data, avoids geometric illusions and detail distortion, and achieves a geometric consistency accuracy of over 98% between the generated result and the target physical entity.

[0054] The data acquisition terminal uses edge-end accuracy verification based on sparse sampling and adopts a hierarchical verification strategy to balance computing power and accuracy.

[0055] The spherical space centered on the target object is divided into multi-resolution grid cells. During acquisition, the number of valid keyframes and view vectors falling within each grid cell are counted in real time. Only when the grid coverage reaches a preset integrity threshold is the next level of verification allowed; otherwise, re-acquisition logic is triggered. After acquisition, the viewpoint, quantity, and quality are verified. Once verification is successful, 3D modeling begins.

[0056] To adapt to the computing power limitations of the acquisition terminal's GPU, instead of directly processing the entire point cloud, an adaptive sparse sampling strategy is implemented. This strategy dynamically adjusts the sampling step size based on the current GPU load of the device, extracting the feature point set that best represents the geometric structure of the object.

[0057] In this embodiment, the specific method for obtaining verification results by performing local accuracy verification on three-dimensional geometric data includes:

[0058] A sparse sampling strategy is used to extract feature points from the three-dimensional geometric data to obtain a feature point set;

[0059] The error is compared between the feature point set and the 3D model reconstructed locally from the visual texture data collected at the same time. If the error is less than the preset truth tolerance threshold, the verification is considered successful.

[0060] The error is calculated using the root mean square (RMSE) error. The true tolerance threshold is preset based on the accuracy requirements of the target application scenario. For example, for consumer product appearance display, the RMSE threshold T can be set to 1 / 500 of the overall object size; for precision industrial parts inspection, it can be set to 1 / 2000. In this embodiment, for general industrial modeling, T = 2.0 mm is set. This threshold is stored in the configuration file of the dedicated acquisition terminal and can be fine-tuned according to the calibration accuracy of the specific sensor.

[0061] When verification fails, augmented reality (AR) dynamically generates visual guidance information. It identifies the spatial coordinates of areas with excessive errors or those that are clearly covered. Utilizing ARKit / ARCore's planar tracking capabilities, these abnormal areas are mapped back to the user's camera preview interface, overlaid on the object as a semi-transparent heatmap, such as a red highlight indicating the missing area. The system tracks the user's device movement in real time; once the user points the camera at the abnormal area and completes the necessary data supplementation, the heatmap immediately changes to a pass status, turning green, achieving a WYSIWYG quality control.

[0062] The data acquisition terminal also performs differentiated data compression and structured encapsulation to provide high signal-to-noise ratio constraint data for generative AI models. Specifically, it identifies normal maps that strongly constrain AI generation and background color maps used only for reference. The system uses lossless or low-compression encoding for normal maps to fully preserve the surface details and lighting logic of objects; and high-compression encoding for background color maps. This allows AI to calculate lighting based on real normal data during rendering, rather than fabricating lighting based on statistical probability, thereby accurately restoring the real material texture of physical entities and guiding AI to generate content that conforms to physical lighting rules. The real data of the collected physical goods is converted into structured data and written into a controlled 3D data structure header file. Physical attributes such as actual size and material reflectivity serve as enhancement parameters for subsequent generative AI model prompts, dynamically feeding data into prompts to control AI-generated images or videos.

[0063] The data acquisition terminal employs hardware-level trusted signatures to ensure the immutability of data after it leaves the terminal, and this is coupled with AI-generated constraints to form end-to-end traceability. The digital signature process for the acquired data by the acquisition terminal specifically includes:

[0064] Calculate the cryptographic hash values ​​of the 3D geometric data and the compressed visual texture data respectively to generate a data fingerprint;

[0065] Within a trusted execution environment, a pre-defined private key is used to digitally sign the data fingerprint, generating a signature list file that conforms to the content source and authenticity alliance specifications or similar standards.

[0066] The three-dimensional geometric data, compressed texture data, metadata configuration table, and signature manifest file are encapsulated into a standard container format to generate the final controlled three-dimensional data structure based on physical truth.

[0067] like Figure 2 The diagram shows the structure of the controlled 3D data structure. The controlled 3D data structure is packaged in ZIP format, with the root file named Controlled_3D_Structure.zip. The internal files and field relationships are as follows:

[0068] The root directory (Controlled_3D_Structure.zip) contains the geometry folder, textures folder, manifest folder, and JSON configuration file. Each file corresponds to a unique hash value, which is stored in the JSON configuration file to realize the integrity verification of the controlled 3D data structure.

[0069] The geometry folder stores the 3D geometric data extracted from sparsely sampled feature points (filenamed point_cloud.ply, in Point Cloud format), including coordinate information, accuracy verification results (RMSE values), and feature point set data.

[0070] The textures folder contains high-resolution normal maps (e.g., named normal_high_res.png) and relatively low-resolution base maps (e.g., named base_low_res.png). The resolutions of both maps conform to a preset scaling ratio (e.g., 2x or higher). Both maps are labeled with the asymmetric compression algorithm type and resolution parameters. The normal maps retain geometric details.

[0071] The manifest folder stores the signature file (filename signed_manifest.json), which contains signature information generated by the ECDSAP-256 algorithm, hash values ​​associated with geometric and texture data, and binding controlled 3D data structure integrity verification information.

[0072] The JSON configuration file (filenamed config.json) records the file paths, unique hash values, data acquisition terminal hardware information (sensor model, NPU model), signature timestamps, and accuracy verification results, which are used for verification and parsing on the cloud server.

[0073] The cloud server verifies the legitimacy of the digital signature on the controlled 3D data structure, confirming that the data has not been tampered with and originates from an authenticated acquisition terminal. High-precision 3D geometric data within the controlled 3D data structure is parsed and converted into spatial condition control signals for the generative AI model's timetable. The data structure of these control signals embeds a verification result identifier for the hardware digital signature, ensuring that the generation process only occurs under the premise that the data is trustworthy. The geometric skeleton of the generated content is locked, ensuring that the generated digital content is completely consistent with the physical truth in size, proportion, and structure. Example of a spatial constraint data structure: This condition can be represented as a JSON object containing a field identifying the signature verification status.

[0074] This invention provides a reliable 3D digital content generation method based on physical truth, which significantly improves the accuracy of geometric consistency between AI-generated content and physical objects, multiplies processing efficiency, and adapts to various scenarios such as physical entity digitization, high-precision digital twin construction, and industrial design. Digitally signing the acquired data not only proves the data's origin but also uses the digital signature information as a key to generate a controlled 3D data structure with the 3D geometric data. This structure serves as a spatial constraint for the generative AI model. The cloud server verifies the signature validity of this constraint before generation can proceed, thus eliminating the possibility of using unverified or tampered data as constraints at the algorithmic logic level.

[0075] Hardware-level signature and hash binding mechanism, combined with end-to-end traceability design, enables trusted 3D digital content to have traceability and anti-tampering capabilities. This ensures the authenticity of the source and data integrity of trusted 3D digital content during transmission in untrusted network environments, prevents man-in-the-middle attacks or data injection, and provides judicial-grade evidentiary value to support scenarios such as copyright protection and compliant circulation of trusted 3D digital content, thus solving the data trust problem.

[0076] Because digital signatures and generative AI models are deeply bound to each other at the data structure level, the copyright of any content generated based on these constraints can be traced back to the data collection source represented by the signing hardware. The signature information includes the hardware identifier and timestamp of the data collection terminal, serving as a link in the chain of evidence and enabling copyright anchoring from data to generated content. This solves the problem of digital asset ownership verification and provides native technological protection for the compliant use of digitized physical entities.

[0077] like Figure 3 As shown, another embodiment of the present invention provides a trusted 3D digital content generation system based on physical truth, comprising: a data acquisition terminal, a cloud server, and a generative artificial intelligence model; the data acquisition terminal is used to acquire 3D geometric data and visual texture data of a target object in real time, compress the visual texture data to obtain compressed visual texture data, perform local precision verification on the 3D geometric data to obtain a verification result, if the verification result is unsuccessful, dynamically generate guidance information, if the verification result is successful, digitally sign the acquired data to obtain digital signature information, generate a controlled 3D data structure based on the verified 3D geometric data, compressed visual texture data, and digital signature information, and send the controlled 3D data structure to the cloud server; the cloud server is used to receive and verify the legality of the digital signature of the controlled 3D data structure, parse the 3D geometric data in the controlled 3D data structure to obtain parsed 3D geometric data, convert the parsed 3D geometric data into spatial constraints, and send the spatial constraints to the generative artificial intelligence model; the generative artificial intelligence model is used to generate content that conforms to physical truth and has a trusted source based on the spatial constraints.

[0078] The acquisition terminal includes a local verification module. The local verification module uses a sparse sampling strategy to extract feature points from the 3D geometric data to obtain a feature point set. The feature point set is then compared with the 3D model reconstructed locally from the visual texture data acquired at the same time. If the error is less than the preset truth tolerance threshold, the verification is considered successful.

[0079] The data acquisition terminal includes a hardware security module, which is used to calculate the encrypted hash values ​​of the three-dimensional geometric data and the compressed visual texture data respectively to generate a data fingerprint; and to digitally sign the data fingerprint using a preset private key in a trusted execution environment to generate digital signature information.

[0080] The execution process of each module can be carried out according to the process steps of the artificial intelligence constraint generation method based on physical truth provided in the first embodiment, and will not be described in detail in this embodiment.

[0081] The artificial intelligence constraint generation system and the artificial intelligence constraint generation method based on physical truth provided in this invention are based on the same inventive concept and have the same beneficial effects, and will not be described again here.

[0082] Another embodiment of the present invention provides an electronic device, which includes a processor, an input device, an output device, and a memory. The processor, the input device, the output device, and the memory are interconnected. The memory is used to store a computer program, which includes program instructions. The processor is configured to call the program instructions to execute the method described in the first embodiment above.

[0083] It should be understood that, in the embodiments of the present invention, the processor may be a Central Processing Unit (CPU), but it may also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0084] Input devices may include touchpads, microphones, etc., while output devices may include displays (LCDs, etc.), speakers, etc.

[0085] The memory may include read-only memory and random access memory, and provides instructions and data to the processor. A portion of the memory may also include non-volatile random access memory. For example, the memory may also store information about the device type.

[0086] In specific implementations, the processor, input device, and output device described in the embodiments of the present invention can execute the implementation of the method embodiments described in the embodiments of the present invention, or they can execute the implementation of the system embodiments described in the embodiments of the present invention, which will not be repeated here.

[0087] The present invention also provides an embodiment of a computer-readable storage medium storing a computer program, the computer program including program instructions, which, when executed by a processor, cause the processor to perform the method described in the first embodiment above.

[0088] The computer-readable storage medium can be an internal storage unit of the terminal described in the foregoing embodiments, such as the terminal's hard drive or memory. The computer-readable storage medium can also be an external storage device of the terminal, such as a plug-in hard drive, Smart Media Card (SMC), Secure Digital (SD) card, or Flash Card equipped on the terminal. Furthermore, the computer-readable storage medium can include both internal storage units and external storage devices of the terminal. The computer-readable storage medium is used to store the computer program and other programs and data required by the terminal. The computer-readable storage medium can also be used to temporarily store data that has been output or will be output.

[0089] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0090] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the terminals and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0091] In the several embodiments provided in this application, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices or units, or may be electrical, mechanical or other forms of connection.

[0092] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for generating reliable 3D digital content based on physical truth, characterized in that, include: The acquisition terminal acquires the three-dimensional geometric data and visual texture data of the target object in real time. The visual texture data is compressed to obtain compressed visual texture data. The three-dimensional geometric data is locally verified to obtain a verification result. If the verification result is unsuccessful, guidance information is dynamically generated. If the verification result is successful, the acquired data is digitally signed to obtain digital signature information. Based on the verified three-dimensional geometric data, compressed visual texture data and digital signature information, a controlled three-dimensional data structure is generated and sent to the cloud server. The cloud server receives and verifies the legality of the digital signature of the controlled 3D data structure, parses the 3D geometric data in the controlled 3D data structure to obtain the parsed 3D geometric data, converts the parsed 3D geometric data into spatial constraints, and sends the spatial constraints to the generative artificial intelligence model. Generative artificial intelligence models generate content that conforms to physical truth and has a reliable source based on spatial constraints.

2. The method according to claim 1, characterized in that, The specific method for obtaining the verification result by performing local accuracy verification on the three-dimensional geometric data includes: A sparse sampling strategy is used to extract feature points from the three-dimensional geometric data to obtain a feature point set; The error is compared between the feature point set and the 3D model reconstructed locally from the visual texture data collected at the same time. If the error is less than the preset truth tolerance threshold, the verification is considered successful.

3. The method according to claim 2, characterized in that, The sparse sampling strategy employs an adaptive dynamic step size, dynamically adjusting the sampling density based on the real-time load status of the graphics processor at the acquisition terminal.

4. The method according to claim 1, characterized in that, The specific methods for digitally signing the acquired data include: Calculate the cryptographic hash values ​​of the 3D geometric data and the compressed visual texture data respectively to generate a data fingerprint; Within a trusted execution environment, a pre-defined private key is used to digitally sign the data fingerprint, generating digital signature information.

5. The method according to claim 1, characterized in that, The acquisition terminal acquires the three-dimensional geometric data of the target object in real time, including anti-counterfeiting liveness detection of the target object, specifically including: Acquire multi-frame 3D geometric data of the target object in a continuous time series and lock the sampling points in a static coordinate system; Calculate the time dimension variance of the sampling point depth values ​​and perform frequency domain analysis on the trend of the time dimension variance; If a periodic peak related to the refresh rate and frequency multiplication factor of the display device is detected in the fluctuation spectrum of the depth value, or if moiré fringe interference features are present, then the target object is determined to be a secondary imaging device that is not a real physical entity, and the generation of the controlled three-dimensional data structure is terminated.

6. A reliable 3D digital content generation system based on physical truth, characterized in that, include: Data acquisition terminals, cloud servers, and generative artificial intelligence models; The acquisition terminal is used to acquire the three-dimensional geometric data and visual texture data of the target object in real time, compress the visual texture data to obtain compressed visual texture data, perform local precision verification on the three-dimensional geometric data to obtain a verification result, if the verification result is unsuccessful, guide information is dynamically generated, if the verification result is successful, digital signature is performed on the acquired data to obtain digital signature information, and a controlled three-dimensional data structure is generated based on the verified three-dimensional geometric data, compressed visual texture data and digital signature information, and the controlled three-dimensional data structure is sent to the cloud server. The cloud server is used to receive and verify the legality of the digital signature of the controlled three-dimensional data structure, parse the three-dimensional geometric data in the controlled three-dimensional data structure to obtain the parsed three-dimensional geometric data, convert the parsed three-dimensional geometric data into spatial constraints, and send the spatial constraints to the generative artificial intelligence model. The generative artificial intelligence model is used to generate content that conforms to physical truth and has a reliable source, based on spatial constraints.

7. The system according to claim 6, characterized in that, The acquisition terminal includes a local verification module. The local verification module uses a sparse sampling strategy to extract feature points from the three-dimensional geometric data to obtain a feature point set. The feature point set is then compared with the three-dimensional model reconstructed locally from the visual texture data acquired at the same time. If the error is less than the preset truth tolerance threshold, the verification is deemed successful.

8. The system according to claim 6, characterized in that, The acquisition terminal includes a hardware security module, which is used to calculate the encrypted hash values ​​of the three-dimensional geometric data and the compressed visual texture data respectively, generate a data fingerprint, and digitally sign the data fingerprint using a preset private key in a trusted execution environment to generate digital signature information.

9. An electronic device, comprising: The processor, input device, output device, and memory are interconnected, the memory being used to store a computer program, the computer program including program instructions, characterized in that the processor is configured to invoke the program instructions to perform the method as described in any one of claims 1-5.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, the computer program including program instructions that, when executed by a processor, cause the processor to perform the method as described in any one of claims 1-5.

Images