Face recognition method and device based on industrial internet of things, medium and electronic equipment

Abstract

The embodiment of the application discloses a face recognition method and device based on an industrial Internet of Things, a medium and an electronic device, relates to the technical field of face recognition, and realizes face recognition by combining two dimensions and three dimensions. On one hand, the reflectivity of skin in the near-infrared wave band is different from that of paper and a screen through infrared image amplification, and the texture features of blood vessel distribution on a real face are extracted. On the other hand, the depth change of facial features is constructed by using 3D point clouds, and the three-dimensional features of the real face are reflected. If either aspect fails, it is considered that the anti-counterfeiting detection fails. Since the front is independent detection, in the case that the single mode anti-counterfeiting detection passes, the multi-modal features are comprehensively utilized, and then a neural network model is used for anti-counterfeiting detection again. The two-dimensional and three-dimensional features are fused to improve the feature quality, more accurate recognition of the face authenticity is realized, and the security of face recognition is improved under multiple safeguards.

Description

Technical Field

[0001] This application relates to the field of facial recognition technology, specifically to a facial recognition method, device, medium, and electronic device based on the Industrial Internet of Things. Background Technology

[0002] Facial recognition is a biometric technology that identifies individuals based on their facial features. It involves using cameras or webcams to capture images or video streams containing faces, automatically detecting and tracking faces within the images, and then performing facial recognition. This process is also commonly referred to as image recognition or facial identification. With the development of the Internet of Things (IoT), existing technologies combine facial recognition with IoT to achieve personnel identification and management in industrial IoT systems, improving management effectiveness. However, this method is vulnerable to external attacks. For example, facial recognition may rely on printed photos or images displayed on other devices. If these flat images are of good quality, their facial features can be recognized, achieving results comparable to those captured by a camera, impacting personnel management and compromising the security of industrial IoT systems and their applications. Summary of the Invention

[0003] The main purpose of this application is to provide a face recognition method, device, medium and electronic device based on the Industrial Internet of Things, which aims to solve the problem of low security of image-based face recognition in the prior art.

[0004] To achieve the above objectives, the technical solutions adopted in the embodiments of this application are as follows:

[0005] In a first aspect, embodiments of this application provide a face recognition method based on the Industrial Internet of Things (IIoT), applied to an IIoT system. The IIoT system includes: a user platform, a service platform, a management platform, a sensor network platform, and an object platform that interact sequentially. The object platform is used to access facial recognition devices. The method includes the following steps:

[0006] Infrared texture features are obtained from the infrared image of the target face; these features are used to characterize the distribution of facial blood vessels.

[0007] Based on the 3D point cloud image of the target face, 3D depth features are obtained; among them, 3D depth features are used to characterize the depth changes of facial features.

[0008] Anti-counterfeiting detection is performed based on infrared images or 3D point cloud images;

[0009] If the anti-counterfeiting detection is passed, the infrared texture features and 3D depth features are stitched together and input into the neural network model to obtain the face detection results.

[0010] In one possible implementation of the first aspect, anti-counterfeiting detection is performed based on infrared images or 3D point cloud images, including:

[0011] The infrared image is binarized to obtain the first infrared image;

[0012] The first infrared image is input into a binary classification model for anti-counterfeiting detection.

[0013] In one possible implementation of the first aspect, binarizing the infrared image to obtain a first infrared image includes:

[0014] Pixels are obtained based on the pixel distribution of the infrared image;

[0015] Using each pixel as the center point and its grayscale value as the center pixel, the size of the neighboring pixels is compared with the center pixel to obtain the binary comparison result.

[0016] Arrange the comparison results in a predetermined order to obtain the number of binary permutations;

[0017] Convert the binary permutation to decimal to obtain the decimal number;

[0018] The center pixel of the center point is updated to a decimal number for binarization to obtain the first infrared image.

[0019] In one possible implementation of the first aspect, anti-counterfeiting detection is performed based on infrared images or 3D point cloud images, including:

[0020] Based on the 3D point cloud image, obtain the target 3D point cloud data; wherein, the target 3D point cloud data represents facial features with depth variations;

[0021] Curvature analysis is performed based on the target 3D point cloud data for anti-counterfeiting detection.

[0022] In one possible implementation of the first aspect, anti-counterfeiting detection is performed based on infrared images or 3D point cloud images, including:

[0023] The first contour features are obtained from the 3D point cloud image;

[0024] The consistency between the first contour feature and the second contour feature is determined for anti-counterfeiting detection; wherein, the second contour feature is obtained based on the RGB image of the target face.

[0025] In one possible implementation of the first aspect, infrared texture features and 3D depth features are concatenated and input into a neural network model to obtain face detection results, including:

[0026] The infrared texture features and 3D depth features are concatenated into a feature vector to obtain a concatenated feature vector.

[0027] The concatenated feature vectors are input into the neural network model to weight and fuse the independent output probabilities of each modality, and then output the face detection results.

[0028] In one possible implementation of the first aspect, before obtaining infrared texture features based on the infrared image of the target face, the method further includes:

[0029] In response to the target face being held within the portrait capture range for a certain period of time, infrared images and 3D point cloud images of the target face are obtained through the portrait acquisition device.

[0030] Secondly, embodiments of this application provide a face recognition device based on the Industrial Internet of Things (IIoT), applied to an IIoT system. The IIoT system includes: a user platform, a service platform, a management platform, a sensor network platform, and an object platform that interact sequentially. The object platform is used to access the face capture device, and the management platform includes:

[0031] The first acquisition module is used to obtain infrared texture features based on the infrared image of the target face; wherein, the infrared texture features are used to characterize the distribution of facial blood vessels;

[0032] The second acquisition module is used to obtain 3D depth features based on the 3D point cloud image of the target face; wherein, the 3D depth features are used to characterize the depth changes of facial features.

[0033] The first detection module is used to perform anti-counterfeiting detection based on infrared images or 3D point cloud images;

[0034] The second detection module is used to stitch together infrared texture features and 3D depth features and input them into a neural network model to obtain face detection results if the anti-counterfeiting detection is passed.

[0035] Thirdly, embodiments of this application provide a computer-readable storage medium storing a computer program, which, when loaded and executed by a processor, implements the face recognition method based on the Industrial Internet of Things as provided in any of the first aspects above.

[0036] Fourthly, embodiments of this application provide an electronic device, including a processor and a memory, wherein,

[0037] Memory is used to store computer programs;

[0038] The processor is used to load and execute computer programs to cause electronic devices to perform the face recognition method based on the Industrial Internet of Things provided in any of the first aspects above.

[0039] Compared with the prior art, the beneficial effects of this application are:

[0040] This application proposes a face recognition method, device, medium, and electronic device based on the Industrial Internet of Things. The method includes: obtaining infrared texture features based on an infrared image of a target face; wherein the infrared texture features are used to characterize the distribution of facial blood vessels; obtaining 3D depth features based on a 3D point cloud image of the target face; wherein the 3D depth features are used to characterize the depth changes of facial features; performing anti-counterfeiting detection based on the infrared image or the 3D point cloud image; and if the anti-counterfeiting detection passes, concatenating the infrared texture features and the 3D depth features and inputting them into a neural network model to obtain a face detection result. This application achieves facial recognition by combining two-dimensional and three-dimensional methods. On the one hand, it can amplify the reflectivity of the skin in the near-infrared band and the difference between it and paper and screen by using infrared images to extract the texture features of blood vessel distribution on real faces. On the other hand, it uses 3D point clouds to construct the depth changes of facial features, reflecting the three-dimensional features of real faces. Failure of either method is considered as failure of anti-counterfeiting detection. Since the previous detections are independent, if the anti-counterfeiting detection of a single modality passes, the multi-modal features are combined and a neural network model is used for a second anti-counterfeiting detection. The fusion of two-dimensional and three-dimensional features improves feature quality and achieves more accurate recognition of the authenticity of faces. Multiple safeguards enhance the security of facial recognition. Attached Figure Description

[0041] Figure 1 This is a schematic diagram of the electronic device structure of the hardware operating environment involved in the embodiments of this application;

[0042] Figure 2 A flowchart illustrating the face recognition method based on the Industrial Internet of Things provided in this application embodiment;

[0043] Figure 3 A schematic diagram of the industrial IoT system in the face recognition method based on industrial IoT provided in the embodiments of this application;

[0044] The diagram is labeled as follows: 101-Processor, 102-Communication bus, 103-Network interface, 104-User interface, 105-Memory. Detailed Implementation

[0045] It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit this application.

[0046] See attached document Figure 1 , attached Figure 1This is a schematic diagram of the electronic device structure of the hardware operating environment involved in the embodiments of this application. The electronic device may include: a processor 101, such as a central processing unit (CPU), a communication bus 102, a user interface 104, a network interface 103, and a memory 105. The communication bus 102 is used to realize the connection and communication between these components. The user interface 104 may include a display screen and an input unit such as a keyboard. Optionally, the user interface 104 may also include a standard wired interface and a wireless interface. The network interface 103 may optionally include a standard wired interface and a wireless interface (such as a Wi-Fi interface). The memory 105 may be a storage device independent of the aforementioned processor 101. The memory 105 may be a high-speed random access memory (RAM) or a stable non-volatile memory (NVM), such as at least one disk storage device. The processor 101 may be a general-purpose processor, including a central processing unit, a network processor, etc., or it may be a digital signal processor, an application-specific integrated circuit, a field-programmable gate array or other programmable logic device, discrete gate or transistor logic device, or discrete hardware component.

[0047] Those skilled in the art will understand that the appendix Figure 1 The structure shown does not constitute a limitation on the electronic device and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0048] As attached Figure 1 As shown, the memory 105, which serves as a storage medium, may include an operating system, a network communication module, a user interface module, and a face recognition device based on the Industrial Internet of Things.

[0049] In the appendix Figure 1 In the electronic device shown, the network interface 103 is mainly used for data communication with the network server; the user interface 104 is mainly used for data interaction with the user; the processor 101 and the memory 105 in this application can be set in the electronic device. The electronic device calls the face recognition device based on the industrial Internet of Things stored in the memory 105 through the processor 101 and executes the face recognition method based on the industrial Internet of Things provided in the embodiment of this application.

[0050] See attached document Figure 2Based on the hardware devices of the foregoing embodiments, embodiments of this application provide a face recognition method based on the Industrial Internet of Things (IIoT), applied to an IIoT system. The IIoT system includes: a user platform, a service platform, a management platform, a sensor network platform, and an object platform that interact sequentially. The object platform is used to access a facial recognition device. The method includes the following steps:

[0051] S10: Obtain infrared texture features based on the infrared image of the target face; whereby the infrared texture features are used to characterize the distribution of facial blood vessels.

[0052] In the specific implementation process, the target face, that is, the image of the person who needs to be recognized, is obtained by taking an infrared image with an infrared camera. Since the reflectivity of real skin is different from that of paper and screen in the near-infrared band (850nm-940nm), the texture features formed by the distribution of blood vessels on the face can be extracted by capturing the infrared reflection characteristics of the infrared camera.

[0053] S20: Obtain 3D depth features from the 3D point cloud image of the target face; whereby the 3D depth features are used to characterize the depth changes of facial features.

[0054] In practice, 3D point cloud images are obtained through scanning cameras. High-precision 3D point cloud data can be generated by projecting laser dot matrix onto the face. This data can reflect the depth changes of facial features in three dimensions and is referred to as 3D depth features.

[0055] S30: Anti-counterfeiting detection based on infrared images or 3D point cloud images.

[0056] In practical implementation, anti-counterfeiting detection refers to verifying whether the image being used for facial recognition is that of a real person, preventing external attacks such as printed photos, videos, and masks. By constructing a multimodal anti-counterfeiting mechanism, infrared anti-counterfeiting can be achieved using infrared images, or 3D structured light anti-counterfeiting can be achieved using 3D point cloud images, depending on the specific circumstances.

[0057] Specifically, infrared anti-counterfeiting involves detecting counterfeit goods based on infrared images or 3D point cloud images, including:

[0058] The infrared image is binarized to obtain the first infrared image;

[0059] The first infrared image is input into a binary classification model for anti-counterfeiting detection.

[0060] In the specific implementation process, the infrared image is binarized to highlight the texture features it carries, and a binary classification model is used for detection, such as SVM or lightweight CNN. These models divide the input samples into two different categories through supervised learning, such as true and false in this application. The identification of the true and false of the face corresponds to the pass and fail of the anti-counterfeiting detection.

[0061] Furthermore, the infrared image is binarized to obtain a first infrared image, including:

[0062] Pixels are obtained based on the pixel distribution of the infrared image;

[0063] Using each pixel as the center point and its grayscale value as the center pixel, the size of the neighboring pixels is compared with the center pixel to obtain the binary comparison result.

[0064] Arrange the comparison results in a predetermined order to obtain the number of binary permutations;

[0065] Convert the binary permutation to decimal to obtain the decimal number;

[0066] The center pixel of the center point is updated to a decimal number for binarization to obtain the first infrared image.

[0067] In practical implementation, to improve detection efficiency, a binarization method is provided. This method transforms the pixel value of each pixel in the image to enhance its robustness to changes in illumination. The grayscale value of each pixel is recorded as its threshold and compared with multiple neighboring pixels. The comparison results are binarized in a predetermined order, such as using a circular neighborhood of pixels. The comparisons are performed clockwise with the center pixel. If a neighboring pixel value is not less than the center pixel value, it is marked as 1; otherwise, it is marked as 0. This results in a string of numbers, recorded as a binary permutation. This string is then converted to decimal and used directly as the binarized value of the center pixel. This local binarization method using neighboring pixels is computationally simple and effectively improves detection and recognition efficiency, enhancing the recognizability of feature parts.

[0068] Under 3D structured light anti-counterfeiting, anti-counterfeiting detection is performed based on infrared images or 3D point cloud images, including:

[0069] Based on the 3D point cloud image, obtain the target 3D point cloud data; wherein, the target 3D point cloud data represents facial features with depth variations;

[0070] Curvature analysis is performed based on the target 3D point cloud data for anti-counterfeiting detection.

[0071] In practice, attacks using paper, video, etc., obviously lack three-dimensional features. To identify attacks with three-dimensional features, curvature analysis is used for anti-counterfeiting detection. After laser dot projection, facial depth features can be obtained, and the target area, i.e., the area with more obvious depth changes, such as the bridge of the nose and eye sockets, is obtained. The point cloud data of this feature area is extracted and recorded as the target 3D point cloud data. By performing curvature analysis on this data, the smoothness of its depth changes is identified. Abnormal smoothness or abrupt changes indicate that the changes are unnatural and do not conform to the characteristics of facial changes, and the anti-counterfeiting detection is considered to fail; otherwise, it passes.

[0072] Under 3D structured light anti-counterfeiting, detection is performed by combining RGB images:

[0073] Anti-counterfeiting detection is performed based on infrared images or 3D point cloud images, including:

[0074] The first contour features are obtained from the 3D point cloud image;

[0075] The consistency between the first contour feature and the second contour feature is determined for anti-counterfeiting detection; wherein, the second contour feature is obtained based on the RGB image of the target face.

[0076] In the specific implementation process, anti-counterfeiting detection is achieved by aligning 3D and 2D images. The 3D point cloud image is vertically projected onto the plane to obtain the first contour feature. Then, the second contour feature is extracted from the RGB image. The two features are subjected to contour consistency detection. If the similarity is greater than the set threshold, it is considered to be consistent and the anti-counterfeiting detection is considered to pass. Otherwise, it fails.

[0077] S40: If the anti-counterfeiting detection passes, the infrared texture features and 3D depth features are stitched together and input into the neural network model to obtain the face detection result.

[0078] In practice, one or two anti-counterfeiting detection methods can be selected. Regardless of the method chosen, the anti-counterfeiting detection is performed independently. If any one of the selected methods fails, the entire anti-counterfeiting detection is considered failed. If all selected methods pass, the detection is considered successful. To enhance the security of the anti-counterfeiting detection and provide an additional layer of protection, infrared texture features and 3D depth features are stitched together to improve the quality of the input features. This is then input into a trained neural network model for anti-counterfeiting detection, yielding the face detection result—the final judgment of whether the target face is genuine. After the anti-counterfeiting detection is completed, faces deemed not to be real are not recognized. Faces deemed to be real are processed normally according to the rules and permissions set by the administrator, such as visitor registration and employee access control.

[0079] Specifically, infrared texture features and 3D depth features are concatenated and input into a neural network model to obtain face detection results, including:

[0080] The infrared texture features and 3D depth features are concatenated into a feature vector to obtain a concatenated feature vector.

[0081] The concatenated feature vectors are input into the neural network model to weight and fuse the independent output probabilities of each modality, and then output the face detection results.

[0082] In practical implementation, the neural network model is trained using samples or historical data acquired through the same methods. The infrared branch corresponds to the infrared texture feature vector, and the 3D branch corresponds to the 3D depth feature vector. The two vectors are concatenated to obtain a concatenated feature vector, which is then input into the multi-branch neural network model. Each modality independently outputs its probability, and the classification result is output through weighted fusion. This can further improve the accuracy of detection. By introducing RGB images into the branches, an additional branch based on the color feature vector acquired from the RGB image is set up to restore the true color of the target and improve the accuracy of anti-counterfeiting detection.

[0083] In one embodiment, before obtaining infrared texture features based on an infrared image of the target face, the method further includes:

[0084] In response to the target face being held within the portrait capture range for a certain period of time, infrared images and 3D point cloud images of the target face are obtained through the portrait acquisition device.

[0085] In the specific implementation process, in order to avoid misidentification and waste of resources, a threshold for holding time is set. The target face needs to be held within the image capture range for a certain period of time, such as adjusting the head to be held within the camera's acquisition range for 2 seconds. After confirming that it is the person who needs to be recognized, the image acquisition device will start collecting data.

[0086] In this embodiment, facial recognition is achieved by combining two-dimensional and three-dimensional methods. On the one hand, infrared images can be used to magnify the difference between the reflectivity of the skin in the near-infrared band and that of paper and screens, extracting the texture features of blood vessel distribution on a real face. On the other hand, 3D point clouds are used to construct the depth changes of facial features, reflecting the three-dimensional features of a real face. Failure of either method is considered a failure of the anti-counterfeiting detection. Since the previous detections were independent, if the anti-counterfeiting detection of a single modality is passed, the multi-modal features are combined and a neural network model is used for another anti-counterfeiting detection. The fusion of two-dimensional and three-dimensional features improves the feature quality and achieves more accurate recognition of the authenticity of the face. Multiple safeguards enhance the security of facial recognition.

[0087] The framework of the industrial IoT system provided in this application embodiment, applied to a face recognition management scenario, is attached. Figure 3 As shown, the architecture can include a user platform, service platform, management platform, sensor network platform, and object platform that interact sequentially, forming a basic five-platform architecture. The object platform can include multiple object sub-platforms for connecting to different facial recognition devices. The sensor network platform can include a device management module and a data transmission management module. The device management module can include a network management unit, a command management unit, and a device status management unit. The data transmission management module can include a data protocol management unit, a data parsing unit, a data classification unit, a data transmission monitoring unit, and a data transmission security unit. The management platform can include a device management sub-platform, a business management sub-platform, and a data center. The device management sub-platform and the business management sub-platform can interact with the data center respectively. The device management sub-platform can include a device operation status monitoring management unit and a metering data monitoring unit. The system comprises a management unit, an equipment parameter management unit, and an equipment lifecycle management unit. The business management sub-platform may include a revenue management unit, a business operator management unit, an application management unit, a message management unit, a scheduling management unit, a sales difference management unit, an operation analysis management unit, and a comprehensive business management unit. The functional modules of the industrial IoT-based facial recognition device can be integrated into the management platform and stored in the data center for retrieval. The service platform may include a management service module, an operation service module, and a security service module. The user platform may include a general user module, a government user module, and a regulatory user module. Through the interaction between the various functional platforms of the industrial IoT system based on the above five platforms, a complete closed-loop information operation logic is established, ensuring the orderly operation of perceived and control information and realizing intelligent management of equipment local operation safety.

[0088] Based on the same inventive concept as in the foregoing embodiments, this application also provides a face recognition device based on the Industrial Internet of Things (IIoT), applied to an IIoT system. The IIoT system includes: a user platform, a service platform, a management platform, a sensor network platform, and an object platform that interact sequentially. The object platform is used to access the face capture device, and the management platform includes:

[0089] The first acquisition module is used to obtain infrared texture features based on the infrared image of the target face; wherein, the infrared texture features are used to characterize the distribution of facial blood vessels;

[0090] The second acquisition module is used to obtain 3D depth features based on the 3D point cloud image of the target face; wherein, the 3D depth features are used to characterize the depth changes of facial features.

[0091] The first detection module is used to perform anti-counterfeiting detection based on infrared images or 3D point cloud images;

[0092] The second detection module is used to stitch together infrared texture features and 3D depth features and input them into a neural network model to obtain face detection results if the anti-counterfeiting detection is passed.

[0093] Those skilled in the art should understand that the division of the various modules in the embodiments is merely a logical functional division. In actual applications, they can be fully or partially integrated into one or more actual carriers. These modules can be implemented entirely in software through processing unit calls, entirely in hardware, or a combination of software and hardware. It should be noted that each module in the face recognition device based on the Industrial Internet of Things in this embodiment corresponds one-to-one with each step in the face recognition method based on the Industrial Internet of Things in the aforementioned embodiments. Therefore, the specific implementation of this embodiment can refer to the implementation of the face recognition method based on the Industrial Internet of Things in the aforementioned embodiments, and will not be repeated here.

[0094] Based on the same inventive concept as in the foregoing embodiments, embodiments of this application also provide a computer-readable storage medium storing a computer program, which, when loaded and executed by a processor, implements the face recognition method based on the Industrial Internet of Things provided in the embodiments of this application.

[0095] Based on the same inventive concept as in the foregoing embodiments, embodiments of this application also provide an electronic device, including a processor and a memory, wherein,

[0096] Memory is used to store computer programs;

[0097] The processor is used to load and execute computer programs to enable electronic devices to perform the face recognition method based on the Industrial Internet of Things provided in the embodiments of this application.

[0098] In some embodiments, the computer-readable storage medium may be a memory such as FRAM, ROM, PROM, EPROM, EEPROM, flash memory, magnetic surface memory, optical disk, or CD-ROM; or it may be a device including one or any combination of the above-mentioned memories. The computer may be a variety of computing devices, including smart terminals and servers.

[0099] In some embodiments, executable instructions may take the form of a program, software, software module, script, or code, written in any form of programming language (including compiled or interpreted languages, or declarative or procedural languages), and may be deployed in any form, including as a standalone program or as a module, component, subroutine, or other unit suitable for use in a computing environment.

[0100] As an example, executable instructions may, but do not necessarily, correspond to files in the file system. They may be stored as part of a file that holds other programs or data, for example, in one or more scripts in a Hyper Text Markup Language (HTML) document, in a single file dedicated to the program in question, or in multiple collaborative files (e.g., a file that stores one or more modules, subroutines, or code sections).

[0101] As an example, executable instructions can be deployed to execute on a single computing device, or on multiple computing devices located in one location, or on multiple computing devices distributed across multiple locations and interconnected via a communication network.

[0102] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0103] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0104] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as read-only memory / random access memory, magnetic disk, optical disk) and includes several instructions to cause a multimedia terminal device (which may be a mobile phone, computer, television receiver, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0105] In summary, this application provides a face recognition method, device, medium, and electronic device based on the Industrial Internet of Things. The method includes: obtaining infrared texture features based on an infrared image of a target face; wherein the infrared texture features are used to characterize the distribution of facial blood vessels; obtaining 3D depth features based on a 3D point cloud image of the target face; wherein the 3D depth features are used to characterize the depth changes of facial features; performing anti-counterfeiting detection based on the infrared image or the 3D point cloud image; and if the anti-counterfeiting detection passes, concatenating the infrared texture features and the 3D depth features and inputting them into a neural network model to obtain a face detection result. This application achieves facial recognition by combining two-dimensional and three-dimensional methods. On the one hand, it can amplify the reflectivity of the skin in the near-infrared band and the difference between it and paper and screen by using infrared images to extract the texture features of blood vessel distribution on real faces. On the other hand, it uses 3D point clouds to construct the depth changes of facial features, reflecting the three-dimensional features of real faces. Failure of either method is considered as failure of anti-counterfeiting detection. Since the previous detections are independent, if the anti-counterfeiting detection of a single modality passes, the multi-modal features are combined and a neural network model is used for a second anti-counterfeiting detection. The fusion of two-dimensional and three-dimensional features improves feature quality and achieves more accurate recognition of the authenticity of faces. Multiple safeguards enhance the security of facial recognition.

[0106] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A face recognition method based on the Industrial Internet of Things, characterized in that, The method is applied to an industrial Internet of Things (IIoT) system, which includes: a user platform, a service platform, a management platform, a sensor network platform, and an object platform that interact sequentially. The object platform is used to access facial recognition devices. The method includes the following steps: Infrared texture features are obtained from an infrared image of a target face; wherein the infrared texture features are used to characterize the distribution of facial blood vessels; prior to obtaining the infrared texture features from the infrared image of the target face, the method further includes: In response to the target face being held within the portrait capture range for a time exceeding a threshold, the infrared image and 3D point cloud image of the target face are obtained through the portrait acquisition device. Based on the 3D point cloud image of the target face, 3D depth features are obtained; wherein, the 3D depth features are used to characterize the depth changes of facial features; Anti-counterfeiting detection is performed based on the infrared image or the 3D point cloud image; the anti-counterfeiting detection based on the infrared image or the 3D point cloud image includes: Based on the 3D point cloud image, target 3D point cloud data is obtained; wherein, the target 3D point cloud data represents facial features with depth variations; Curvature analysis is performed on the target 3D point cloud data for anti-counterfeiting detection; The anti-counterfeiting detection based on the infrared image or the 3D point cloud image includes: Based on the 3D point cloud image, the first contour feature is obtained; The consistency between the first contour feature and the second contour feature is determined for anti-counterfeiting detection; wherein, the second contour feature is obtained based on the RGB image of the target face; If the anti-counterfeiting detection passes, the infrared texture features and the 3D depth features are concatenated and input into a neural network model to obtain a face detection result; the concatenation of the infrared texture features and the 3D depth features and input into the neural network model to obtain a face detection result includes: The infrared texture features and the 3D depth features are concatenated to obtain a concatenated feature vector; The concatenated feature vector is input into a neural network model to weight and fuse the independent output probabilities of each modality, and then output the face detection result.

2. The face recognition method based on the Industrial Internet of Things according to claim 1, characterized in that, The anti-counterfeiting detection based on the infrared image or the 3D point cloud image includes: The infrared image is binarized to obtain a first infrared image; The first infrared image is input into a binary classification model for anti-counterfeiting detection.

3. The face recognition method based on the Industrial Internet of Things according to claim 2, characterized in that, The step of binarizing the infrared image to obtain a first infrared image includes: Pixels are obtained based on the pixel distribution of the infrared image; Taking each pixel as the center point and its gray value as the center pixel, the size of the neighboring pixels is compared with the center pixel to obtain a binary comparison result; Arrange the comparison results in a predetermined order to obtain the number of binary permutations; Convert the binary permutation number to decimal to obtain a decimal number; The center pixel of the center point is updated to the decimal number for binarization to obtain the first infrared image.

4. A facial recognition device based on the Industrial Internet of Things, characterized in that, This is applied to an industrial Internet of Things (IIoT) system, which includes: a user platform, a service platform, a management platform, a sensor network platform, and an object platform that interact sequentially. The object platform is used to access facial recognition devices, and the management platform includes: The first acquisition module is used to obtain infrared texture features based on the infrared image of the target face; wherein the infrared texture features are used to characterize the distribution of facial blood vessels; before obtaining the infrared texture features based on the infrared image of the target face, the module further includes: In response to the target face being held within the portrait capture range for a time exceeding a threshold, the infrared image and 3D point cloud image of the target face are obtained through the portrait acquisition device. The second acquisition module is used to obtain 3D depth features based on the 3D point cloud image of the target face; wherein the 3D depth features are used to characterize the depth changes of facial features. The first detection module is used to perform anti-counterfeiting detection based on the infrared image or the 3D point cloud image; the anti-counterfeiting detection based on the infrared image or the 3D point cloud image includes: Based on the 3D point cloud image, target 3D point cloud data is obtained; wherein, the target 3D point cloud data represents facial features with depth variations; Curvature analysis is performed on the target 3D point cloud data for anti-counterfeiting detection; The anti-counterfeiting detection based on the infrared image or the 3D point cloud image includes: Based on the 3D point cloud image, the first contour feature is obtained; The consistency between the first contour feature and the second contour feature is determined for anti-counterfeiting detection; wherein, the second contour feature is obtained based on the RGB image of the target face; The second detection module is used to, upon successful anti-counterfeiting detection, concatenate the infrared texture features and the 3D depth features, and input the result into a neural network model to obtain a face detection result; the concatenation of the infrared texture features and the 3D depth features, and input into the neural network model to obtain a face detection result, includes: The infrared texture features and the 3D depth features are concatenated to obtain a concatenated feature vector; The concatenated feature vector is input into a neural network model to weight and fuse the independent output probabilities of each modality, and then output the face detection result.

5. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is loaded and executed by the processor, it implements the face recognition method based on the Industrial Internet of Things as described in any one of claims 1-3.

6. An electronic device, characterized in that, Including processor and memory, among which, The memory is used to store computer programs; The processor is used to load and execute the computer program to enable the electronic device to perform the face recognition method based on the Industrial Internet of Things as described in any one of claims 1-3.

Images