Method and system for generating designated standing posture prompt image, and millimeter wave self-service scanning security inspection method and system

Abstract

Provided in the present disclosure are a method and system for generating a designated standing posture prompt image, and a millimeter wave self-service scanning security inspection method and system. The method for generating a designated standing posture prompt image comprises: acquiring a standing posture image of a subject by means of a camera device; on the basis of the standing posture image of the subject, acquiring a body shape and standing posture image of the subject; on the basis of a standard standing posture image, adjusting the body shape and standing posture image of the subject, acquiring a standing posture prompt image of the subject, and using the standing posture prompt image as a designated standing posture prompt image of the subject; on the basis of the standing posture image of the subject obtained by means of the camera device and body posture features, acquiring a current standing posture image of the subject; and simultaneously displaying the designated standing posture prompt image and the current standing posture image of the subject to the subject, wherein the designated standing posture prompt image may be different according to body posture features of different subjects.

Description

Methods and systems for generating designated standing posture prompts, and methods and systems for millimeter-wave self-service scanning security checks.

[0001] This disclosure claims priority to Chinese Patent Application No. 202411998320.0, filed on December 31, 2024, the contents of which are incorporated herein by reference. Technical Field

[0002] This disclosure relates to the field of security inspection, specifically image generation during a security inspection process, and more specifically, to a system and method for generating a specified standing posture image, as well as a millimeter-wave self-service scanning security inspection system and method including the system for generating the specified standing posture image. Background Technology

[0003] The background description provided herein is intended to present the general context of this disclosure. To the extent described in this background section, the work of the currently identified inventors and aspects of the description that may not constitute related art at the time of filing are neither explicitly nor implicitly considered to be related art to this disclosure.

[0004] Millimeter-wave body imaging technology is currently an advanced technology in the global security field and is used in airports worldwide for passenger security checks. This device can effectively detect items concealed under clothing on various parts of the body without direct contact, especially non-metallic items, and can extract information such as the shape, size, and location of concealed items from the image. Furthermore, millimeter-wave body imaging equipment is harmless to the human body and has strong penetrating power; its transmission power is less than one-thousandth of the electromagnetic radiation from a mobile phone, accurately identifying items carried by the body, effectively improving the objectivity, accuracy, and targeting of inspections, reducing the workload of security personnel, and increasing security efficiency.

[0005] Due to the limitations of the imaging principle, when being scanned by a millimeter-wave security scanner in the security checkpoint, the person being inspected must assume a specific posture. Their arms and legs must be spread out at a certain angle; otherwise, obstructions may occur, such as arms blocking the armpits or legs blocking the inner thighs, or artifacts may form between the legs and under the armpits. Furthermore, the Automatic Target Recognition (ATR) module also has posture requirements when identifying suspects and converting the original image into a cartoon display. If an incorrect posture is detected, the person being inspected will be required to adjust their posture and be rescanned. When the equipment operator or guide is unfamiliar with the posture requirements, frequent posture errors will occur during scanning, thus requiring significant time for training.

[0006] Therefore, there is an urgent need for a method and system that can generate specified standing posture prompts that are adapted to different examinees. Summary of the Invention

[0007] This disclosure provides a method for generating a designated standing posture prompt image. The method includes: obtaining a standing posture image of a subject through a camera device; obtaining a body shape standing posture image of the subject based on the standing posture image; adjusting the body shape standing posture image of the subject based on a standard standing posture image, and obtaining a standing posture prompt image of the subject, and using the standing posture prompt image as the designated standing posture prompt image of the subject; obtaining a current standing posture image of the subject based on the standing posture image and body characteristics obtained through the camera device; and simultaneously displaying the designated standing posture prompt image and the current standing posture image of the subject to the subject, wherein the designated standing posture prompt image can vary according to the body characteristics of different subjects.

[0008] In one embodiment of this disclosure, obtaining the body shape and standing posture image of the examinee includes: obtaining the key point coordinates of the examinee based on the standing posture image of the examinee, and locating the key points of the examinee's body using a key point localization model. The key point localization model is based on a convolutional neural network model, which includes multiple residual blocks, convolutional layers, pooling layers, and fully connected layers.

[0009] In one embodiment of this disclosure, the key points of the human body include the head, shoulders, elbows, wrists, hips, knees, and ankles.

[0010] In one embodiment of this disclosure, obtaining the body shape and standing posture diagram of the examinee includes: obtaining the body parts corresponding to the located key points of the human body, and connecting the obtained body parts to obtain the outline of the examinee's current standing posture diagram.

[0011] In one embodiment of this disclosure, obtaining the body shape and standing posture diagram of the examinee further includes: calculating the translation distance and / or scaling factor of the key points based on the standard standing posture diagram and key points of the human body, and performing translation and scaling operations on the obtained body shape and standing posture diagram based on the translation distance and / or scaling factor.

[0012] In one embodiment of this disclosure, obtaining the subject's current standing posture image includes: obtaining the subject's current standing posture image using a human body 3D reconstruction model, wherein the human body 3D reconstruction model includes binocular camera registration processing, point cloud stitching and noise reduction processing, point cloud planar mapping processing, and GPU acceleration processing.

[0013] In one embodiment of this disclosure, the method for generating a specified standing posture prompt image further includes: comparing the current standing posture image of the examinee with the specified standing posture prompt image, and displaying the comparison result to the examinee through a display device to guide the examinee to place their current standing posture image within the range of the specified standing posture prompt image.

[0014] In one embodiment of this disclosure, the method for generating a specified standing posture prompt image further includes: when the subject's previous standing posture image is within the range of the specified standing posture prompt image, the specified standing posture prompt image is displayed in a first color; when at least a portion of the subject's previous standing posture image is outside the range of the specified standing posture prompt image, the specified standing posture prompt image is displayed in a second color, the second color being different from the first color.

[0015] In one embodiment of this disclosure, the method for generating a specified standing posture prompt image further includes: generating a specified standing posture prompt image of a fixed size, scaling the current standing posture image to make it the same size as the specified standing posture prompt image, and displaying it together with the specified standing posture prompt image to the examinee.

[0016] This disclosure also provides a millimeter-wave security inspection method, which includes the following steps: generating a designated standing posture prompt image according to the method described in any of the above embodiments; continuously capturing images of the subject using a camera device; determining whether the subject's current standing posture is consistent with the designated standing posture prompt image; when the subject's current standing posture is inconsistent with the designated standing posture prompt image, reminding the subject to adjust their standing posture; when the subject's current standing posture is consistent with the designated standing posture prompt image, inspecting the subject using a millimeter-wave security inspection instrument and identifying and marking the location corresponding to the suspect.

[0017] This disclosure also provides a system for generating a designated standing posture prompt image, comprising: a camera device configured to acquire image data of a subject's standing posture and body shape; a processing module configured to obtain a body shape standing posture image of the subject based on the subject's standing posture image; adjust the subject's body shape standing posture image based on a standard standing posture image to obtain a standing posture prompt image of the subject, and use the standing posture prompt image as the designated standing posture prompt image of the subject; obtain the subject's current standing posture image based on the subject's standing posture image and body shape characteristics; and simultaneously display the designated standing posture prompt image and the current standing posture image of the subject to the subject, wherein the designated standing posture prompt image varies according to the body shape characteristics of different subjects.

[0018] In one embodiment of this disclosure, the specified standing posture generation system further includes a display panel for displaying the current standing posture diagram and the specified standing posture prompt diagram.

[0019] In one embodiment of this disclosure, the camera device includes a visible light camera device, an infrared camera device, and a binocular depth camera device.

[0020] In one embodiment of this disclosure, the running processing module includes a key point localization model for determining the key point coordinates of a subject's standing posture diagram. The key point localization model is based on a convolutional neural network model, which includes multiple residual blocks, convolutional layers, pooling layers, and fully connected layers.

[0021] In one embodiment of this disclosure, the running processing module is configured to obtain the key point coordinates of the examinee based on the examinee's standing posture diagram, obtain the human body parts corresponding to the key points of the human body according to the located human body key points, connect the obtained human body parts to obtain the outline of the examinee's current standing posture diagram, and calculate the translation distance and / or scaling factor of the key points according to the standard standing posture diagram and human body key points, and perform translation and scaling operations on the obtained body shape standing posture diagram based on the translation distance and / or scaling factor.

[0022] In one embodiment of this disclosure, the operation processing module is further configured to simultaneously display the current standing posture diagram and the specified standing posture prompt diagram to the examinee to guide the examinee to adjust the current standing posture. Adjusting the standing posture includes adjusting the raising and / or lowering of the arms and the closing and / or separating of the legs.

[0023] This disclosure also provides a millimeter-wave security inspection system, wherein the millimeter-wave security inspection system includes a designated posture generation system, a millimeter-wave security inspection device, and an inspection channel as described in any of the systems above, the millimeter-wave security inspection device being configured to perform millimeter-wave scanning inspection on the person being inspected to identify and mark suspicious objects, and the inspection channel being configured for the person being inspected to perform millimeter-wave scanning inspection thereon.

[0024] These and other aspects of this disclosure will become apparent from the following description of preferred embodiments in conjunction with the accompanying drawings and description, but variations and modifications may be made thereto without departing from the spirit and scope of the novel concept of this disclosure. Attached Figure Description

[0025] This disclosure will be more fully understood from the detailed description and accompanying drawings. These drawings illustrate one or more embodiments of this disclosure and, together with the written description, serve to explain the principles of this disclosure. Where possible, the same reference numerals are used throughout the drawings to denote the same or similar elements of the embodiments, and wherein:

[0026] Figure 1 is a flowchart of a method for generating a specified standing posture diagram according to an exemplary embodiment of the present disclosure.

[0027] Figure 2 is a schematic diagram of the standing posture of a subject according to an exemplary embodiment of the present disclosure.

[0028] Figure 3 is a diagram of the current standing posture of the examinee and a diagram indicating a specified standing posture according to an exemplary embodiment of the present disclosure.

[0029] Figure 4 is a three-dimensional reconstruction model of the human body according to an exemplary embodiment of the present disclosure.

[0030] Figure 5 is a standard standing posture diagram and a body shape standing posture diagram of an examinee according to an exemplary embodiment of the present disclosure.

[0031] Figure 6 is a flowchart of a millimeter-wave self-service scanning security inspection method according to an exemplary embodiment of the present disclosure.

[0032] Figure 7 is a schematic diagram of artifacts between the arms and torso and between the legs of different subjects according to an exemplary embodiment of the present disclosure.

[0033] Figure 8 is a schematic diagram of a millimeter-wave self-service scanning security inspection system according to an exemplary embodiment of the present disclosure. Detailed Implementation

[0034] The present disclosure will now be described more fully with reference to the accompanying drawings, which illustrate exemplary embodiments of the present disclosure. However, the present disclosure may be implemented in various ways and should not be construed as limited to the embodiments described herein. These embodiments are provided to make the disclosure more thorough and complete, and to fully convey the scope of the disclosure to those skilled in the art. In the drawings, the thickness and area of ​​layers may be enlarged for clarity. Throughout the specification, the same reference numerals are used to denote the same elements. For different embodiments, elements may have different relationships and different positions.

[0035] This disclosure proposes a method for generating a specified standing posture prompt image. This method can be applied to locations requiring security checks, such as airports and train stations, and can generate a specific standing posture prompt image for each person being checked. In other words, different standing posture prompt images are generated based on the physical characteristics of each person being checked, such as height, weight, etc. Because a specific standing posture prompt image is generated for each person being checked, security check operations can be performed more accurately and conveniently.

[0036] As shown in Figure 1, the method for generating a designated standing posture prompt image includes the following steps: In step S1, a standing posture image of the examinee is obtained through a camera device; in step S2, a body shape standing posture image of the examinee is obtained through the standing posture image of the examinee; in step S3, the body shape standing posture image of the examinee is adjusted based on the standard standing posture image, and a standing posture prompt image of the examinee is obtained, and the standing posture prompt image is used as the designated standing posture prompt image of the examinee; in step S4, the current standing posture image of the examinee is obtained based on the standing posture image and body characteristics of the examinee obtained through the camera device; in step S5, the designated standing posture prompt image and the current standing posture image of the examinee are displayed to the examinee simultaneously, wherein the designated standing posture prompt image can be different according to the body characteristics of different examinees.

[0037] In one embodiment of this disclosure, when the examinee is located in the inspection channel, an image of the examinee can be captured using a camera device, preferably capturing the examinee's standing posture image, as shown on the right side of Figure 2. The processing module 103 uses a key point localization model to process and obtain the key point coordinates of the examinee, and then uses the key point coordinates of the examinee's standing posture image to calculate the examinee's graphic standing posture image.

[0038] The keypoint localization model described is based on a convolutional neural network model with an improved residual network. This convolutional neural network model can include: N residual blocks, convolutional layers, pooling layers, and fully connected layers. In a preferred embodiment, a ResNet convolutional neural network with 3 residual blocks can be used as the backbone network; however, this disclosure does not limit the number of residual blocks, which can be more or less than 3. Each residual block has a basically the same structure, consisting of two parts: one part is a 3x3 convolutional layer processed using batch normalization and ReLU activation functions; the other part is a 1x1 convolutional layer. Finally, the two parts are summed and processed using batch normalization and ReLU activation functions. The residual block first passes through a 3x3 convolutional layer with a stride of 2, followed by a 3x3 convolutional layer with a stride of 1, reducing the number of channels. The structure of the residual block effectively reduces the number of parameters and computational cost of convolution. The batch normalization algorithm also accelerates network convergence and improves training speed.

[0039] In this embodiment, the processing module 103 locates the key points of the subject's body using a key point localization model. This model is based on a convolutional neural network (CNN) model, but is not limited to VGG or ResNet models; the most suitable model can be selected based on actual needs. Furthermore, the key point localization model also supports integrated models formed by multiple models such as VGG and ResNet to improve localization accuracy and robustness. The model selection mechanism automatically selects the most suitable key point localization model based on different application scenarios and requirements to ensure optimal performance in various environments. The key points include at least the head, shoulders, elbows, wrists, hips, knees, and ankles.

[0040] In embodiments of this disclosure, as shown in Figures 4 and 5, the processing module 103 uses key point coordinate information to calculate the body shape and standing posture diagram of the examinee. The processing module 103 can detect important key points of the human body based on the standing posture positioning model; it processes the key points corresponding to different parts of the human body using appropriate difference or fitting algorithms to generate contour curves or line segments for each part; then, according to the structural order of the human body, it splices these contour curves or line segments together to form a complete body shape and standing posture diagram. In a preferred embodiment, the extracted key points are also sorted and arranged according to the logical order of human body parts, for example, head first, then neck, torso, and finally limbs. After arranging the key point set, any erroneous or abnormal key points are removed.

[0041] In embodiments of this disclosure, as shown in Figures 4 and 5, the translation distance and / or scaling factor of the key points are calculated based on the standard standing posture diagram and key human body points. The obtained body posture diagram is then translated and scaled based on the translation distance and / or scaling factor. The processing module 103 can determine the translation distance based on key points at the center of the human body, such as the center point of the shoulder, to align the subject's body posture diagram with the standard standing posture diagram in a planar position. Size measurements (e.g., height, shoulder width) are calculated from the key points, and the corresponding scaling factor is calculated. Then, the subject's body posture diagram is enlarged or reduced according to the scaling factor to match the standard standing posture in overall size or local details (e.g., key point areas), generating the final human standing posture prompt diagram.

[0042] Finally, in the embodiments of this disclosure, as shown in Figures 3, 6, and 8, the processing module 103 can compare the examinee's current standing posture diagram with the designated standing posture prompt diagram, and display the comparison result to the examinee through a display device. This guides the examinee to adjust their current standing posture according to the designated standing posture prompt diagram to meet its requirements. For example, adjusting the raising and / or lowering of the arms, the closing and / or opening of the legs, or the angle between the extended arms and the open legs. The designated standing posture prompt diagram varies depending on the examinee's body shape. For example, the designated standing posture prompt diagram can generate different prompt images based on the examinee's height, weight, or build.

[0043] In embodiments of this disclosure, the processing module 103 can also generate a projected image of the subject's current standing posture based on the subject's posture and body features captured by the camera device, using a human body 3D reconstruction model. The current standing posture projection image is a continuously calculated, real-time generated image. In an exemplary embodiment of this disclosure, the human body 3D reconstruction model may include binocular camera registration processing, point cloud stitching and denoising processing, point cloud planar mapping processing, and GPU-accelerated processing.

[0044] In a preferred embodiment, the main function of the binocular camera registration technology is to register the two cameras to the world coordinate system coordinate axes XYZ. For example, the registered three-dimensional space takes the orientation of the human body in the correct standing posture as the Z-axis, and the XY plane as the front of the human body in the correct standing posture.

[0045] In a preferred embodiment, the purpose of point cloud stitching technology is to stitch together the single-sided human point clouds obtained after registering the cameras on both sides into a complete human point cloud. Point cloud denoising technology can remove noise from the point cloud to achieve a better human body display effect. The point cloud stitching algorithm includes a coarse point cloud registration algorithm and a point cloud fusion algorithm. In an exemplary embodiment, the algorithm first extracts key points from the point cloud data on both sides of the human body, and uses the coarse point cloud registration algorithm and the key points to calculate the rigid body transformation matrix of the point cloud. Then, the registered point clouds are clustered, filtered, and weighted to obtain the stitched complete human point cloud. Finally, Gaussian filtering and SOR denoising algorithms are used to denoise the point cloud.

[0046] In a preferred embodiment, the main function of the point cloud-plane mapping technique is to map the obtained 3D human point cloud model onto a 2D plane for display and subsequent algorithmic use. A frontal image of a human in a correct standing posture can be the XY plane in a coordinate system. This technique maps the human point cloud onto the XY plane while retaining the data on the Z-axis, so that subsequent detection tasks can be performed on the 2D image while simultaneously obtaining the world coordinates on the Z-axis based on the pixel coordinates.

[0047] In a preferred embodiment, all algorithms used in the human body 3D reconstruction technology are written as GPU kernel operators and processed on the GPU to ensure the real-time performance of the algorithms.

[0048] In one embodiment of this disclosure, the method for generating a specified standing posture prompt image further includes: generating a specified standing posture prompt image of a fixed size, scaling the current standing posture image to make it the same size as the specified standing posture prompt image, and displaying it together with the specified standing posture prompt image to the examinee.

[0049] Inside the inspection channel, the inspection module generates a fixed-size posture prompt image. The camera continuously captures images of the person being inspected. The inspection module then generates a current posture image and scales it to the same size as the specified posture prompt image. These images are then displayed together (e.g., projected onto an inner channel panel using a projector). Simultaneously, it identifies whether the person's posture meets the requirements of the specified posture prompt image. In other words, if the person's current posture image does not meet the requirements of the specified posture prompt image, a color-coded prompt is given; for example, the outline of the specified posture prompt image is displayed in red (as shown in the left image of Figure 3), prompting the person to change their posture. If the person's posture meets the requirements of the specified posture prompt image, the outline of the specified posture prompt image is displayed in green (as shown in the right image of Figure 3). As shown in Figure 8, the millimeter-wave security scanner performs millimeter-wave scanning inspections on the person being inspected and identifies and marks suspicious items. The person being inspected can undergo millimeter-wave scanning inspections within the inspection channel.

[0050] As shown in Figure 6, in another embodiment of this disclosure, a flowchart of a millimeter-wave self-service scanning security inspection method is also disclosed, which includes the following steps:

[0051] S10: During the period when the person being inspected 102 enters the inspection channel 109 defined by the millimeter wave security scanner, the camera device 104 acquires image data of the person being inspected 102, such as taking a photo or video of the person being inspected 102, and transmits it to the operation processing module 103.

[0052] S11: The operation processing module 103 calculates the physical characteristics of the examinee based on the image data of the examinee captured by the camera device;

[0053] S12: The operation processing module 103 uses the method for generating a specified standing posture prompting diagram in this disclosure to generate a specified standing posture prompting diagram corresponding to the characteristics of the examinee 102.

[0054] S13: The processing module 103 displays the specified standing posture prompt diagram and the current standing posture diagram to the examinee 102 through the display device 110.

[0055] S14: The camera device 104 continuously captures images or videos of the subject 102 and sends the image data to the operation processing module 103.

[0056] S15: The running processing module 103 determines whether the current standing posture of the examinee 102 is consistent with the specified standing posture prompt diagram.

[0057] S16: If the current standing posture matches the specified standing posture prompt image, the millimeter-wave security scanner scans and inspects the person 102. For example, if the current standing posture image falls completely within the outline of the specified standing posture prompt image, the person's current standing posture is considered to meet the requirements (as shown on the right side of Figure 3).

[0058] S17: If the current standing posture does not match the designated standing posture prompt diagram, guide the examinee 102 to adjust their posture via image and / or voice. For example, if a part of the examinee extends beyond the outline of the designated standing posture prompt diagram, provide the examinee with a corresponding reminder to change their posture and adjust according to the outline of the designated standing posture prompt diagram until the projection of the examinee's current standing posture is completely within the outline of the designated standing posture prompt diagram. Adjustments to the examinee's standing posture may include, but are not limited to, adjusting the raising and / or lowering of the arms, and the bringing and / or separating of the legs.

[0059] S18: The millimeter-wave security scanner performs millimeter-wave scanning on the person being inspected and marks the location of the suspected items carried by the person being inspected 102.

[0060] As shown in Figure 8, this disclosure also proposes a system for generating a designated standing posture prompt image. The system includes: a camera device configured to acquire image data of the standing posture and body shape of a subject; a processing module configured to obtain the key point coordinates of the subject's standing posture image from the subject's standing posture image, and calculate the subject's body shape standing posture image using the key point coordinates; adjust the subject's body shape standing posture image based on a standard standing posture image, and obtain a standing posture prompt image for the subject, using the prompt image as the designated standing posture prompt image for the subject; obtain the subject's current standing posture image using a three-dimensional human body reconstruction model based on the subject's standing posture image and body shape characteristics; and simultaneously display the designated standing posture prompt image and the current standing posture image to the subject, wherein the designated standing posture prompt image varies according to the body shape characteristics of different subjects.

[0061] This designated posture generation system can be applied in a millimeter-wave detector 100 to perform millimeter-wave security detection on the subject 102. The camera device 104 of the designated posture generation system may include a visible light camera, an infrared camera, or a binocular depth camera. The designated posture generation system may also include a projector 105. Furthermore, the designated posture generation system may include a display panel 106 for displaying the current posture diagram and the designated posture prompt diagram.

[0062] In one embodiment of this disclosure, the processing module can generate key point coordinate information of the examinee in real time, and compare and analyze the detected key point coordinate information with the corresponding standard in the standard standing posture, so as to better guide the examinee to adjust the standing posture.

[0063] In one embodiment of this disclosure, the processing module 103 can calculate the angle α between the line L connecting the i-th key point of the examinee and the adjacent (i+1)-th key point, and the line L' connecting the i-th key point and the adjacent (i-1)-th key point. The angle α between the obtained line L and line L' is compared with the corresponding angle β between key points in the standard standing posture. Based on the comparison result, the angle α between the obtained line L and line L' is adjusted so that its difference from the angle β between the corresponding key points in the standard standing posture is within a certain range. For example, the angle between α and β is set to not exceed γ degrees, where γ is 10 and i is a positive integer greater than 1. That is, the angle error between the line connecting the key points in the standing posture model and the line connecting the key points in the standard standing posture can be controlled within 10° as a basis for determining whether the posture conforms to the standard standing posture.

[0064] Figure 5 (left side) shows a schematic diagram of the angles connecting the key points of the standing posture standard. As shown, the left shoulder point is designated as the i-th key point, referred to as key point b for clarity. The (i-1)-th key point is designated as key point a, which is the left elbow point. The line connecting the left elbow point and the left shoulder point is line segment ab. The (i+1)-th key point is designated as key point c, which is the left chest point. The line connecting key point b (left shoulder point) and key point c (left chest point) is line segment bc. For example, the angle β between key point connecting lines ab and bc is 45°. Figure 5 (right side) shows a schematic diagram of the key point connecting lines for the current standing posture. Key points a, b, and c correspond to key points d, e, and f, respectively. That is, key point d is the left elbow point, key point e is the left shoulder point, and key point f is the left chest point. In the standard standing posture, the angle α between the key point connecting line ef and the key point connecting line de is 15°. The difference between the angle in the current standing posture and the angle in the standard standing posture exceeds the set value of γ by 10°. Therefore, it can be determined that the left hand part in the standing posture on the right side of Figure 5 is unqualified, while the other positions are qualified, thus forming information to adjust the left hand posture.

[0065] In the method for generating a specified standing posture cue image proposed in this disclosure, as shown in Figure 7, artifacts between the arms and torso, as well as between the legs, are reduced as the angle increases; therefore, a larger angle is desirable. However, excessively large angles of arm extension and leg spread, coupled with the requirement to maintain the posture for 2 seconds, can cause discomfort for the examinee. Furthermore, fixed footprint cues within the passageway may result in too small a leg angle for taller individuals and too large an angle for shorter individuals, causing discomfort. Therefore, generating a unique standing posture that does not affect the millimeter-wave image while ensuring the examinee's comfort is crucial for each examinee of different body types.

[0066] First, by generating customized standing postures based on individual body types, millimeter-wave signals can ensure full-body coverage, reducing detection blind spots caused by improper posture. For example, adjusting arm angles ensures effective signal coverage of the waist and thighs. Second, inappropriate standing postures can lead to artifacts or false alarms during scanning; the system may misjudge ordinary clothing wrinkles or body postures as potential threats. Personalized standing posture prompts can significantly reduce false alarm rates and improve overall detection efficiency. Third, allowing examinees to adjust their standing posture according to their own circumstances avoids the discomfort caused by forced uniform postures. For example, shorter individuals may feel uncomfortable in certain postures, while customized standing posture guidance allows examinees to cooperate more naturally during the inspection. Finally, during peak security check periods, customized standing postures can speed up the scanning process, reduce adjustment time, and allow the system to adapt more quickly to the scanning needs of different groups of people.

[0067] The terminology used herein is for illustrative purposes only and should not be construed as limiting the meaning or scope of this disclosure. As used herein, the singular form may include the plural form unless a specific example is clearly indicated in the context. Furthermore, the expressions “comprising” and / or “including” as used herein do not limit the shapes, numbers, steps, operations, components, elements, and / or groups thereof mentioned, nor do they exclude the appearance or inclusion of one or more other different shapes, numbers, steps, operations, components, elements, and / or groups thereof, or inclusion thereof.

[0068] As used herein, terms such as “first,” “second,” etc., are used to describe various components, assemblies, regions, and / or parts. These terms are used only to distinguish one component, assembly, region, layer, or part from another component, assembly, region, or part. Therefore, the description of a first component, assembly, region, layer, or part may also refer to a second component, assembly, region, or part without departing from the scope of this disclosure.

[0069] The foregoing description of exemplary embodiments of this disclosure is for illustrative and descriptive purposes only and is not intended to be exhaustive or to limit this disclosure to the precise forms disclosed. Many modifications and variations are possible in accordance with the foregoing teachings. The embodiments were chosen and described to explain the principles of this disclosure and its practical application, so that others skilled in the art can utilize this disclosure and various embodiments with various modifications suitable for the particular purpose contemplated. Alternative embodiments will become apparent to those skilled in the art to which this disclosure pertains without departing from the spirit and scope of this disclosure. Therefore, the scope of this disclosure is defined by the appended claims rather than by the foregoing description and the exemplary embodiments described therein.

Claims

1. A method for generating a cue image for a specified standing posture, comprising: A standing posture image of the examinee is obtained through a camera device; A body shape standing posture diagram of the examinee is obtained by using the standing posture diagram described by the examinee; Based on the standard standing posture diagram, the body shape standing posture diagram of the examinee is adjusted, and the standing posture prompt diagram of the examinee is obtained, and the standing posture prompt diagram is used as the designated standing posture prompt diagram of the examinee. Based on the subject's standing posture and body characteristics obtained through a camera device, the subject's current standing posture is obtained; and The designated standing posture prompt image and the current standing posture image are simultaneously displayed to the examinee. The specified standing posture prompt diagram can vary depending on the physical characteristics of different examinees.

2. The method according to claim 1, wherein, The process of obtaining the subject's body shape and posture image includes: obtaining the key point coordinates of the subject based on the subject's posture image, and using a key point localization model to locate the key points of the subject's body. The key point localization model is based on a convolutional neural network model, which includes multiple residual blocks, convolutional layers, pooling layers, and fully connected layers.

3. The method according to claim 2, wherein, The key points of the human body include the head, shoulders, elbows, wrists, hips, knees, and ankles.

4. The method according to claim 3, wherein, The process of obtaining the subject's body shape and standing posture diagram includes: obtaining the body parts corresponding to the located key points of the human body, and connecting the obtained body parts to obtain the outline of the subject's current standing posture diagram.

5. The method according to claim 4, wherein, The process of obtaining the body shape and standing posture diagram of the examinee further includes: calculating the translation distance and / or scaling factor of the key points based on the standard standing posture diagram and key points of the human body, and performing translation and scaling operations on the obtained body shape and standing posture diagram based on the translation distance and / or scaling factor.

6. The method according to claim 1, wherein, The process of obtaining the subject's current standing posture image includes: obtaining the subject's current standing posture image using a human body 3D reconstruction model, wherein the human body 3D reconstruction model includes binocular camera registration processing, point cloud stitching and noise reduction processing, point cloud planar mapping processing, and GPU acceleration processing.

7. The method according to claim 1, wherein, The method for generating a specified standing posture prompt image further includes: comparing the subject's current standing posture image with the specified standing posture prompt image, and displaying the comparison result to the subject through a display device to guide the subject to position their current standing posture image within the range of the specified standing posture prompt image.

8. The method according to claim 1, wherein, The method for generating a specified standing posture prompt image further includes: when the subject's previous standing posture image is within the range of the specified standing posture prompt image, the specified standing posture prompt image is displayed in a first color; when at least a portion of the subject's previous standing posture image is outside the range of the specified standing posture prompt image, the specified standing posture prompt image is displayed in a second color, the second color being different from the first color.

9. The method according to claim 1, wherein, The method for generating a specified standing posture prompt image further includes: generating a specified standing posture prompt image of a fixed size, scaling the current standing posture image to make it the same size as the specified standing posture prompt image, and displaying it together with the specified standing posture prompt image to the examinee.

10. A millimeter-wave self-service scanning security inspection method, comprising: The method according to any one of claims 1-9 generates a specified standing posture prompt image; Continuously capture images of the examinee using a camera device; Determine whether the examinee's current standing posture matches the specified standing posture prompt image; When the examinee's current standing posture is inconsistent with the designated standing posture prompt diagram, the examinee is reminded to adjust their standing posture; When the examinee's current standing posture matches the designated standing posture prompt, the millimeter-wave security scanner is used to inspect the examinee and identify and mark the location of the suspected object.

11. A system for generating a specified standing posture prompt image, comprising: The camera device is configured to acquire image data of the subject's posture and body shape; The processing module is configured to obtain the subject's body type standing posture diagram from the subject's standing posture diagram; adjust the subject's body type standing posture diagram based on the standard standing posture diagram, and obtain the subject's standing posture prompt diagram, and use the standing posture prompt diagram as the subject's designated standing posture prompt diagram; obtain the subject's current standing posture diagram based on the subject's standing posture diagram and body characteristics; and simultaneously display the subject's designated standing posture prompt diagram and the current standing posture diagram to the subject, wherein the designated standing posture prompt diagram varies according to the body characteristics of different subjects.

12. The specified standing posture generation system according to claim 11, wherein, The specified standing posture generation system also includes a display panel for displaying the current standing posture diagram and the specified standing posture prompt diagram.

13. The specified standing posture generation system according to claim 11, wherein, The camera device includes a visible light camera device, an infrared camera device, and a binocular depth camera device.

14. The specified stance generation system according to any one of claims 11-13, wherein, The operation processing module includes a key point localization model for determining the key point coordinates of the subject's standing posture diagram. The key point localization model is based on a convolutional neural network model, which includes multiple residual blocks, convolutional layers, pooling layers, and fully connected layers.

15. The specified stance generation system according to any one of claims 11-13, wherein, The processing module is configured to obtain the coordinates of key points of the examinee based on the examinee's standing posture diagram, obtain the corresponding human body parts based on the located human body key points, connect the obtained human body parts to obtain the outline of the examinee's current standing posture diagram, and calculate the translation distance and / or scaling factor of the key points based on the standard standing posture diagram and human body key points, and perform translation and scaling operations on the obtained body shape standing posture diagram based on the translation distance and / or scaling factor.

16. The designated stance generation system according to any one of claims 11-13, wherein, The operation processing module is also configured to simultaneously display the current standing posture diagram and the specified standing posture prompt diagram to the examinee to guide the examinee to adjust the current standing posture. Adjusting the standing posture includes adjusting the raising and / or lowering of the arms and the closing and / or separating of the legs.

17. A millimeter-wave self-service scanning security inspection system, wherein, The millimeter-wave security inspection system includes a designated posture generation system as described in any one of claims 11-16, a millimeter-wave security scanner, and an inspection channel, wherein the millimeter-wave security scanner is configured to perform millimeter-wave scanning inspection on the person being inspected to identify and mark suspicious objects, and the inspection channel is configured for the person being inspected to undergo millimeter-wave scanning inspection thereon.

Images