Method and device for evaluating authenticity of 3D digital assets and medium

Abstract

The invention relates to the technical field of data processing, in particular to a method and equipment for evaluating the authenticity of 3D digital assets and a storage medium, and aims to solve the technical problem of how to intuitively and accurately evaluate the authenticity of the 3D digital assets in synthetic data. In order to achieve the purpose, the method comprises the steps of obtaining first point cloud data of a to-be-evaluated 3D digital asset, obtaining second point cloud data of a real 3D target having the same target category as the to-be-evaluated 3D digital asset, obtaining a comparison similarity between the first point cloud data and the second point cloud data by applying a trained neural network model, and obtaining the 3D digital asset to be evaluated according to the comparison similarity. According to the method, the authenticity evaluation result of the to-be-evaluated 3D digital assets is obtained, the authenticity evaluation of the 3D digital assets based on the point cloud data can be realized, the accuracy of the evaluation process is improved, the subjective error of human eye judgment can be avoided, and the evaluation process has better consistency. And the evaluation process does not depend on a scene, so that the accuracy of the evaluation process can be further improved.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a method, device and storage medium for authenticity assessment of 3D digital assets. Background Art

[0002] Synthetic data is data artificially created through algorithms or models. Synthetic data can amplify the value of real data and effectively solve the pain points of difficult data collection, expensive labeling, and slow iteration. It can not only support model training based on simulation learning, but also build a new paradigm of training and evaluation through closed-loop simulation. Synthetic data can be applied to multiple fields such as autonomous driving.

[0003] In the process of synthetic data production, it is often necessary to evaluate the 3D digital assets in the synthetic data to determine the authenticity of the 3D digital assets, thereby further ensuring the availability of the synthetic data. In the prior art, two methods are generally used to evaluate the authenticity of 3D digital assets in synthetic data.

[0004] One is to achieve authenticity evaluation based on 2D projection of 3D digital assets. That is, use various means to find the mapping or projection of 3D digital assets in 2D, and then make authenticity comparisons at the 2D level. However, since 3D digital assets include not only surface material information, but also the size and shape information of the asset itself, simply looking at the 2D projection in many cases cannot discover the detailed differences in shape and size, and aligning 2D is actually a very challenging task.

[0005] Another method is to use AR (Augmented Reality) or VR (Virtual Reality) to compare the authenticity. That is, put the 3D digital asset into a real scene, and then let the human eye subjectively judge whether it is real. This situation is often affected by the subjective influence of the human eye, and the consistency of the evaluation process is poor. At the same time, judging the authenticity of the asset from the scene environment is often affected by the scene itself.

[0006] Accordingly, the art needs a new 3D digital asset authenticity assessment solution to solve the above problems. Summary of the invention

[0007] In order to overcome the above-mentioned defects, the present application is proposed to solve or at least partially solve the technical problem of how to intuitively and accurately evaluate the authenticity of 3D digital assets in synthetic data.

[0008] In a first aspect, a method for evaluating the authenticity of a 3D digital asset is provided, the method comprising: Obtain the first point cloud data of the 3D digital asset to be evaluated; Acquire second point cloud data of a real 3D target having the same target category as the 3D digital asset to be evaluated; Based on the trained neural network model, obtaining a comparison similarity between the first point cloud data and the second point cloud data according to the first point cloud data and the second point cloud data; According to the comparison similarity, an authenticity evaluation result of the 3D digital asset to be evaluated is obtained.

[0009] In a technical solution of the authenticity evaluation method of the above 3D digital asset, the obtaining, based on the trained neural network model and according to the first point cloud data and the second point cloud data, the comparison similarity between the first point cloud data and the second point cloud data comprises: Based on the trained neural network model, respectively obtaining a first point cloud feature of the first point cloud data and a second point cloud feature of the second point cloud data; The comparison similarity is obtained according to the first point cloud feature and the second point cloud feature.

[0010] In a technical solution of the authenticity evaluation method of the 3D digital asset, obtaining the comparison similarity according to the first point cloud feature and the second point cloud feature includes: Obtaining a cosine distance between the first point cloud feature and the second point cloud feature; The comparison similarity is obtained according to the cosine distance.

[0011] In a technical solution of the authenticity assessment method of the above 3D digital asset, there are a plurality of the first point cloud features and a plurality of the second point cloud features respectively; The obtaining the cosine distance between the first point cloud feature and the second point cloud feature includes: Respectively acquiring a first cosine distance between each of the first point cloud features and the corresponding second point cloud feature, thereby obtaining a plurality of first cosine distances; The cosine distance is obtained according to an average value of multiple first cosine distances.

[0012] In a technical solution of the authenticity evaluation method of the above 3D digital asset, obtaining the authenticity evaluation result of the 3D digital asset to be evaluated according to the comparison similarity includes: Comparing the comparison similarity with a preset similarity threshold; If the comparison similarity is greater than or equal to the similarity threshold, determining that the authenticity evaluation result is passed; If the comparison similarity is less than the similarity threshold, the authenticity evaluation result is determined to be failed.

[0013] In a technical solution of the authenticity evaluation method of the above 3D digital asset, the step of obtaining the second point cloud data of a real 3D target having the same target category as the 3D digital asset to be evaluated includes: A 3D object is reconstructed according to the real 3D object to obtain the second point cloud data.

[0014] In a technical solution of the above-mentioned 3D digital asset authenticity assessment method, the method further includes training the neural network model according to the following steps: Point cloud data samples of real 3D targets containing multiple target types are used to train the preset neural network model to obtain a trained neural network model.

[0015] In a technical solution of the authenticity assessment method of the above 3D digital asset, the first point cloud data includes color point cloud data; The second point cloud data includes color point cloud data; The point cloud data samples of the real 3D object include color point cloud data samples.

[0016] In one technical solution of the above-mentioned 3D digital asset authenticity assessment method, the neural network model is a point cloud classification model.

[0017] In a second aspect, an electronic device is provided, comprising at least one processor; and a memory communicatively connected to the at least one processor; wherein a computer program is stored in the memory, and when the computer program is executed by the at least one processor, the method described in any one of the technical solutions of the above-mentioned 3D digital asset authenticity assessment method is implemented.

[0018] In a third aspect, a computer-readable storage medium is provided, wherein a plurality of program codes are stored therein, wherein the program codes are suitable for being loaded and run by a processor to execute the method described in any one of the technical solutions of the above-mentioned 3D digital asset authenticity assessment method.

[0019] The above one or more technical solutions of the present application have at least one or more of the following beneficial effects: In implementing the technical solution of the authenticity evaluation method of 3D digital assets provided by this application, this application obtains the first point cloud data of the 3D digital asset to be evaluated, and obtains the second point cloud data of the real 3D target with the same target category as the 3D digital asset to be evaluated, and applies the trained neural network model to obtain the comparative similarity of the first point cloud data and the second point cloud data, and obtains the authenticity evaluation result of the 3D digital asset to be evaluated based on the comparative similarity. Through the above configuration, this application can realize the authenticity evaluation of 3D digital assets based on point cloud data. Point cloud data can reflect the material, size, shape and other information of 3D digital assets, and can better reflect more details of 3D digital assets, thereby improving the accuracy of the evaluation process. At the same time, based on the trained neural network model to achieve the comparative similarity of the first point cloud data and the second point cloud data, it can avoid the subjective error of human eye judgment, so that the evaluation process has better consistency. And the evaluation process of this application does not depend on the scene, which can further improve the accuracy of the evaluation process.

[0020] Furthermore, the first point cloud data and the second point cloud data of the present application both include color point cloud data, and the neural network model is also obtained by training based on color point cloud data samples of real 3D targets. Since the color point cloud data not only contains the three-dimensional spatial coordinates of the point cloud, but also contains color information (such as RGB values), it can provide a variety of three-dimensional information such as the shape, texture, and color of the target, and can more accurately and effectively obtain the comparison similarity between the first point cloud data and the second point cloud data, thereby obtaining a more accurate authenticity evaluation result of the 3D digital asset to be evaluated. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The disclosure of the present application will become easier to understand with reference to the accompanying drawings. It is easy for those skilled in the art to understand that these drawings are only for illustrative purposes and are not intended to limit the scope of protection of the present application. Among them: Figure 1 It is a schematic flow chart of the main steps of a method for authenticity assessment of a 3D digital asset according to an embodiment of the present application; Figure 2 It is a schematic flow chart of the main steps of a method for evaluating the authenticity of a 3D digital asset according to an implementation of an embodiment of the present application; Figure 3 It is a schematic diagram of the main structure of an electronic device according to an embodiment of the present application.

[0022] Reference numerals: 11: memory; 12: processor. DETAILED DESCRIPTION

[0023] Some embodiments of the present application are described below with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are only used to explain the technical principles of the present application and are not intended to limit the protection scope of the present application.

[0024] In the description of the present application, "module" and "processor" may include hardware, software or a combination of the two. A module may include hardware circuits, various suitable sensors, communication ports, memory, and may also include software parts, such as program code, or a combination of software and hardware. The processor may be a central processing unit, a microprocessor, an image processor, a digital signal processor or any other suitable processor. The processor has data and / or signal processing functions. The processor may be implemented in software, hardware or a combination of the two. Computer-readable storage media include any suitable medium that can store program code, such as a disk, a hard disk, an optical disk, a flash memory, a read-only memory, a random access memory, etc. The term "A and / or B" means all possible combinations of A and B, such as only A, only B or A and B. The term "at least one A or B" or "at least one of A and B" has a similar meaning to "A and / or B" and may include only A, only B or A and B. The singular terms "one" and "the" may also include plural forms.

[0025] The relevant user personal information that may be involved in the various embodiments of this application is strictly in accordance with the requirements of laws and regulations, following the principles of legality, legitimacy and necessity, based on the reasonable purposes of business scenarios, to process the personal information that users actively provide during the use of products / services or generated due to the use of products / services, as well as the personal information obtained with the user's authorization.

[0026] The user personal information processed by this application will vary depending on the specific product / service scenario, and shall be based on the specific scenario in which the user uses the product / service, and may involve the user's account information, device information, driving information, vehicle information or other related information. This application will treat the user's personal information and its processing with a high degree of diligence.

[0027] This application attaches great importance to the security of user personal information and has taken reasonable and feasible security protection measures that meet industry standards to protect user information and prevent personal information from being accessed, disclosed, used, modified, damaged or lost without authorization.

[0028] Here we first explain some terms involved in this application.

[0029] Color point cloud refers to point cloud data obtained through 3D scanning technology. Each point contains not only 3D spatial coordinates (usually x, y, z values), but also color information (such as RGB values). This point cloud data can be used to reconstruct the appearance and shape of objects and is widely used in computer vision, virtual reality, autonomous driving and other fields. It can help analyze the shape, texture and color of objects and provide richer 3D information.

[0030] The point cloud classification model is a neural network model used to analyze and identify objects in point cloud data. It extracts features from each point in the point cloud and learns the features of different categories (such as buildings, cars, trees, etc.) to achieve automatic classification.

[0031] 3D reconstruction: 3D reconstruction refers to the process of converting two-dimensional images or point cloud data into a three-dimensional model. It reconstructs the three-dimensional shape and structure of an object or scene by analyzing images or sensor data from multiple perspectives.

[0032] See attached Figure 1 , Figure 1 FIG. 1 is a flow chart of the main steps of a method for evaluating the authenticity of a 3D digital asset according to an embodiment of the present application. Figure 1 As shown, a 3D digital asset authenticity assessment method in an embodiment of the present application mainly includes the following steps S101 to S104.

[0033] Step S101: Acquire the first point cloud data of the 3D digital asset to be evaluated.

[0034] In this embodiment, the first point cloud data of the 3D digital asset to be evaluated can be obtained, wherein the 3D digital asset is a virtual target placed in a synthetic data scene, and may include a dynamic target and a static target.

[0035] Step S102: Acquire second point cloud data of a real 3D target having the same target category as the 3D digital asset to be evaluated.

[0036] In this embodiment, second point cloud data of a real 3D target having the same target category as the 3D digital asset to be evaluated may be obtained.

[0037] In one implementation, the real 3D object may be reconstructed to obtain second point cloud data of the real 3D object.

[0038] In one implementation, a real 3D target may be reconstructed based on methods such as nerf (Neural Radiance Fields) and 3DGS (Gaussian Splatting).

[0039] Step S103: Based on the trained neural network model, the comparison similarity between the first point cloud data and the second point cloud data is obtained according to the first point cloud data and the second point cloud data.

[0040] In this embodiment, the first point cloud data and the second point cloud data may be input into a trained neural network model to obtain a comparative similarity between the first point cloud data and the second point cloud data.

[0041] In one implementation, the neural network model may be a detection model, a segmentation model, etc.

[0042] In one implementation, the neural network model may be a point cloud classification model, and the point cloud classification model may be a model composed of a model architecture such as PointNet, PointNet++, etc.

[0043] In one implementation, point cloud data samples of real 3D targets including multiple target types may be used to train a neural network model to obtain a trained neural network model.

[0044] In one implementation, step S103 may further include the following steps S1031 and S1032: Step S1031: Based on the trained neural network model, first point cloud features of the first point cloud data and second point cloud features of the second point cloud data are respectively obtained.

[0045] In this embodiment, feature extraction of point cloud data may be performed based on the trained neural network model to respectively obtain first point cloud features of the first point cloud data and second point cloud features of the second point cloud data.

[0046] Step S1032: Obtaining a comparison similarity based on the first point cloud feature and the second point cloud feature.

[0047] In this implementation, step S1032 may further include the following steps S10321 and S10322: Step S10321: Obtain the cosine distance between the first point cloud feature and the second point cloud feature.

[0048] Step S10322: Obtain the comparison similarity according to the cosine distance.

[0049] In this embodiment, the first point cloud feature and the second point cloud feature can be converted into a one-dimensional feature vector, and the cosine distance between the feature vector corresponding to the first point cloud feature and the feature vector corresponding to the second point cloud feature can be obtained, and the comparison similarity is obtained according to the cosine distance.

[0050] In one embodiment, when there are multiple first point cloud features and multiple second point cloud features, the first cosine distance between the feature vector corresponding to each first point cloud feature and the feature vector corresponding to the corresponding second point cloud feature can be obtained respectively, and the multiple first cosine distances are averaged to obtain the cosine distance between the first point cloud feature and the second point cloud feature.

[0051] Among them, other methods commonly used in the art for measuring the similarity between feature vectors may also be used to obtain the comparative similarity between the first point cloud data and the second point cloud data.

[0052] Step S104: obtaining the authenticity evaluation result of the 3D digital asset to be evaluated according to the comparison similarity.

[0053] In this embodiment, the authenticity evaluation result of the 3D digital asset to be evaluated can be obtained based on the comparison similarity and the preset similarity threshold.

[0054] In one implementation, step S104 may further include the following steps S1041 to S1043: Step S1041: Compare the comparison similarity with a preset similarity threshold.

[0055] Step S1042: If the comparison similarity is greater than or equal to the similarity threshold, the authenticity evaluation result is determined to be passed.

[0056] Step S1043: If the comparison similarity is less than the similarity threshold, the authenticity evaluation result is determined to be failed.

[0057] In this embodiment, the authenticity evaluation result of the 3D digital asset to be evaluated can be determined based on the comparison result between the comparison similarity and the similarity threshold. If the comparison similarity is greater than or equal to the similarity threshold, the authenticity evaluation result is determined to be passed; if the comparison similarity is less than the similarity threshold, the authenticity evaluation result is determined to be failed. Among them, those skilled in the art can set the value of the similarity threshold according to the needs of actual applications.

[0058] Based on the method described in the above steps S101 to S104, the embodiment of the present application obtains the first point cloud data of the 3D digital asset to be evaluated, and obtains the second point cloud data of the real 3D target of the same target category as the 3D digital asset to be evaluated, and applies the trained neural network model to obtain the comparative similarity of the first point cloud data and the second point cloud data, and obtains the authenticity evaluation result of the 3D digital asset to be evaluated based on the comparative similarity. Through the above configuration, the embodiment of the present application can realize the authenticity evaluation of the 3D digital asset based on the point cloud data. The point cloud data can reflect the material, size, shape and other information of the 3D digital asset, and can better reflect more details of the 3D digital asset, thereby improving the accuracy of the evaluation process. At the same time, based on the trained neural network model to achieve the comparative similarity of the first point cloud data and the second point cloud data, it can avoid the subjective error of human eye judgment, so that the evaluation process has better consistency. And the evaluation process of the embodiment of the present application does not depend on the scene, which can further improve the accuracy of the evaluation process.

[0059] In one embodiment, the first point cloud data may include color point cloud data; the second point cloud data may include color point cloud data; and the point cloud data samples of the real 3D target may include color point cloud data samples. Since both the first point cloud data and the second point cloud data include color point cloud data, and the neural network model is also trained based on the color point cloud data samples of the real 3D target, the color point cloud data not only includes the three-dimensional spatial coordinates of the point cloud, but also includes color information (such as RGB values), which can provide a variety of three-dimensional information such as the shape, texture, and color of the target, and can more accurately and effectively obtain the comparison similarity between the first point cloud data and the second point cloud data, thereby obtaining a more accurate authenticity evaluation result of the 3D digital asset to be evaluated.

[0060] In one embodiment, please refer to the attached Figure 2 , Figure 2 FIG. 1 is a flow chart of the main steps of a method for evaluating the authenticity of a 3D digital asset according to an implementation of an embodiment of the present application. Figure 2 As shown, the authenticity evaluation method of 3D digital assets may include the following steps S201 to S207: Step S201: Use point cloud data samples of real 3D targets containing multiple target types to train a point cloud classification model.

[0061] Step S202: Acquire color point cloud data of the 3D digital asset to be evaluated.

[0062] Step S203: obtaining the first point cloud feature of the 3D digital asset to be evaluated according to the trained point cloud classification model.

[0063] Step S204: Acquire color point cloud data of the real 3D target.

[0064] Step S205: Obtain a second point cloud feature of the real 3D target according to the trained point cloud classification model.

[0065] Step S206: Obtaining a comparison similarity based on the first point cloud feature and the second point cloud feature.

[0066] Step S207: Obtain the authenticity evaluation result of the 3D digital asset to be evaluated according to the comparison similarity.

[0067] It should be pointed out that although the various steps in the above embodiments are described in a specific order, those skilled in the art can understand that in order to achieve the effect of the present application, different steps do not have to be executed in such an order, and they can be executed simultaneously (in parallel) or in other orders. These adjusted schemes are equivalent to the technical schemes described in this application, and therefore will also fall within the scope of protection of this application.

[0068] It is understood by those skilled in the art that all or part of the processes in the method for implementing the above-mentioned embodiment of the present application can also be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable storage medium may include: any entity or device, medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory, random access memory, electrical carrier signal, telecommunication signal and software distribution medium, etc. that can carry the computer program code.

[0069] Another aspect of the present application also provides a computer-readable storage medium.

[0070] In an embodiment of a computer-readable storage medium according to the present application, the computer-readable storage medium may be configured to store a program for executing the authenticity evaluation method of a 3D digital asset of the above method embodiment, and the program may be loaded and run by a processor to implement the authenticity evaluation method of the above 3D digital asset. For ease of explanation, only the parts related to the embodiment of the present application are shown. For specific technical details not disclosed, please refer to the method part of the embodiment of the present application. The computer-readable storage medium may be a storage device formed by various electronic devices. Optionally, the computer-readable storage medium in the embodiment of the present application is a non-temporary computer-readable storage medium.

[0071] Another aspect of the present application provides an electronic device.

[0072] In an embodiment of an electronic device according to the present application, the electronic device may include at least one processor; and a memory connected to the at least one processor; wherein a computer program is stored in the memory, and when the computer program is executed by the at least one processor, the method described in any of the above embodiments is implemented. The electronic device described in the present application may be, but is not limited to, a mobile phone, a tablet computer, a desktop, a laptop, a handheld computer, a notebook computer, a vehicle-mounted device, an ultra-mobile personal computer (UMPC), a netbook, a personal digital assistant (PDA), an augmented reality (AR) device, a virtual reality (VR) device, etc., and the embodiments of the present application are not limited to this. See the attached Figure 3 , Figure 3 FIG. 4 exemplarily shows that the memory 11 and the processor 12 are communicatively connected via a bus.

[0073] So far, the technical solution of the present application has been described in conjunction with an embodiment shown in the accompanying drawings, but it is easy for those skilled in the art to understand that the protection scope of the present application is obviously not limited to these specific embodiments. Without departing from the principles of the present application, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present application.

Claims

1. A method for evaluating the authenticity of a 3D digital asset, characterized in that: The method comprises: Obtain the first point cloud data of the 3D digital asset to be evaluated; Acquire second point cloud data of a real 3D target having the same target category as the 3D digital asset to be evaluated; Based on the trained neural network model, obtaining a comparison similarity between the first point cloud data and the second point cloud data according to the first point cloud data and the second point cloud data; According to the comparison similarity, an authenticity evaluation result of the 3D digital asset to be evaluated is obtained.

2. The authenticity evaluation method of 3D digital assets according to claim 1, characterized in that: The step of obtaining the comparison similarity between the first point cloud data and the second point cloud data based on the trained neural network model according to the first point cloud data and the second point cloud data includes: Based on the trained neural network model, respectively obtaining a first point cloud feature of the first point cloud data and a second point cloud feature of the second point cloud data; The comparison similarity is obtained according to the first point cloud feature and the second point cloud feature.

3. The authenticity evaluation method of 3D digital assets according to claim 2, characterized in that: The obtaining the comparison similarity according to the first point cloud feature and the second point cloud feature includes: Obtaining a cosine distance between the first point cloud feature and the second point cloud feature; The comparison similarity is obtained according to the cosine distance.

4. The authenticity evaluation method of 3D digital assets according to claim 3, characterized in that: There are a plurality of the first point cloud features and a plurality of the second point cloud features respectively; The obtaining the cosine distance between the first point cloud feature and the second point cloud feature includes: Respectively acquiring a first cosine distance between each of the first point cloud features and the corresponding second point cloud feature, thereby obtaining a plurality of first cosine distances; The cosine distance is obtained according to an average value of multiple first cosine distances.

5. The authenticity evaluation method of 3D digital assets according to claim 1, characterized in that: The obtaining the authenticity evaluation result of the 3D digital asset to be evaluated according to the comparison similarity includes: Comparing the comparison similarity with a preset similarity threshold; If the comparison similarity is greater than or equal to the similarity threshold, determining that the authenticity evaluation result is passed; If the comparison similarity is less than the similarity threshold, the authenticity evaluation result is determined to be failed.

6. The authenticity evaluation method of 3D digital assets according to claim 1, characterized in that: The step of obtaining second point cloud data of a real 3D target having the same target category as the 3D digital asset to be evaluated includes: A 3D object is reconstructed according to the real 3D object to obtain the second point cloud data.

7. The authenticity evaluation method of 3D digital assets according to claim 1, characterized in that: The method further comprises training the neural network model according to the following steps: Point cloud data samples of real 3D targets containing multiple target types are used to train the preset neural network model to obtain a trained neural network model.

8. The authenticity evaluation method of 3D digital assets according to claim 7, characterized in that: The first point cloud data includes color point cloud data; The second point cloud data includes color point cloud data; The point cloud data samples of the real 3D object include color point cloud data samples.

9. The authenticity evaluation method of a 3D digital asset according to any one of claims 1 to 8, characterized in that: The neural network model is a point cloud classification model.

10. An electronic device, characterized in that: include: at least one processor; and, a memory communicatively coupled to the at least one processor; The memory stores a computer program, and when the computer program is executed by the at least one processor, the authenticity assessment method of a 3D digital asset according to any one of claims 1 to 9 is implemented.

11. A computer-readable storage medium storing a plurality of program codes, characterized in that: The program code is suitable for being loaded and run by a processor to execute the authenticity assessment method of a 3D digital asset according to any one of claims 1 to 9.

Images