Smart store management and control system for internet bar

Abstract

The invention discloses a smart store management and control system for an internet bar, and relates to the technical field of identity authentication. The system comprises an image acquisition module used for acquiring continuous multi-frame first face images of a target user through a camera in response to the fact that a target verification mode of the target user is a face verification mode; the sequence construction module is used for performing facial feature detection on each first face image and constructing a facial feature missing sequence of the target user; the facial feature missing sequence comprises first facial features which are severely missing in continuous multiple frames of first facial images; the camera adjusting module is used for adjusting the shooting position of the camera based on each first facial feature in the facial feature missing sequence; and the user control module is used for controlling the entering of the target user through a second face image of the target user acquired by the camera after the shooting position is adjusted. According to the invention, the accuracy of identity recognition can be improved.

Description

Technical Field

[0001] This invention relates to the field of identity authentication technology, and more specifically to a smart store management system for internet cafes. Background Technology

[0002] The internet cafe industry faces challenges such as high labor costs, low operational efficiency, and complex management, making the introduction of intelligent management systems a growing trend. With the development of technologies like the Internet of Things (IoT), artificial intelligence (AI), and big data, smart store management systems have become a significant driving force for the industry's growth. These systems, centered on intelligence and automation, propel the industry towards greater refinement and intelligence, with smart access control systems being a crucial component.

[0003] Currently, smart access control systems in internet cafes often use cameras for facial recognition to identify individuals and manage personnel. When a user chooses facial verification, a camera at a fixed location captures the user's facial image. This image is then used for identification and verification, thereby controlling the user's entry.

[0004] However, due to the fixed position of the camera, when the target user is performing facial recognition, factors such as not looking directly at the camera, frequent head movements, and changes in lighting may lead to poor quality facial images captured by the camera, resulting in serious loss of facial features and thus poor accuracy in identity recognition. Summary of the Invention

[0005] This invention provides a smart store management system for internet cafes, which can improve the accuracy of identity recognition.

[0006] A first aspect of this invention provides a smart store management system for internet cafes, comprising: The image acquisition module is used to acquire multiple consecutive frames of the first face image of the target user in response to the target user's target verification method being face verification. The sequence construction module is used to detect facial features in each first face image and construct a facial feature missing sequence for the target user; the facial feature missing sequence includes first facial features that are severely missing in multiple consecutive frames of first face images; The camera adjustment module is used to adjust the shooting position of the camera based on each first facial feature in the missing facial feature sequence. The user control module is used to control the entry of the target user by using the second facial image of the target user obtained by the camera after adjusting the shooting position.

[0007] Furthermore, the present invention also proposes a sequence construction module, comprising: The feature detection unit is used to detect facial features in each first face image and determine the degree of feature loss of each facial feature in each first face image. The probability determination unit is used to determine the probability of severe missing facial features based on the degree of feature loss in each first face image. The sequence construction unit is used to sort each first facial feature with a severe missing probability greater than a first preset probability threshold according to the order of the severe missing probability, so as to obtain the facial feature missing sequence of the target user.

[0008] Furthermore, the present invention also proposes a feature detection unit for: Obtain the current number of features of the second facial feature, the area of ​​the first feature of the second facial feature, and the area of ​​the second feature of the reference facial feature in the target first face image; the target first face image is any first face image, the second facial feature is any facial feature, and the reference facial features are all facial features except the second facial feature; The degree of feature loss of the second facial features in the target first face image is determined by using the ratio between the current number of features of the second facial features and the standard number of features of the second facial features, as well as the average difference between the area of ​​the first feature and the area of ​​each second feature.

[0009] Furthermore, the present invention also proposes a probability determination unit for: The average value of the feature loss degree of the second facial feature in each first face image is obtained by averaging the feature loss degree of the second facial feature; the second facial feature is any one facial feature. The feature loss degree of the second facial features in each first face image is fitted in chronological order to obtain a linear curve of the feature loss degree of the second facial features. The probability of severe loss of the second facial feature is determined by using the average value of the feature loss degree and the slope of the feature loss degree line.

[0010] Furthermore, the present invention also proposes a camera adjustment module, comprising: The probability comparison unit is used to compare the probability of severe loss of the nose feature with a second preset probability threshold in response to each first facial feature including the nose feature, and to obtain the probability comparison result. The voice prompt unit is used to initiate a voice prompt to the target user in response to a probability comparison result indicating that the probability of severe loss of nasal features is greater than a second preset probability threshold, so as to instruct the target user to look directly at the camera. The camera adjustment unit is used to adjust the shooting position of the camera based on the distribution of feature points of the nasal features, in response to the probability comparison result indicating that the probability of severe loss of nasal features is not greater than a second preset probability threshold.

[0011] Furthermore, the present invention also proposes a camera adjustment unit, comprising: The region division subunit is used to divide the nasal feature region corresponding to the nasal feature into a first region to the left of the central feature point and a second region to the right of the central feature point, based on the central feature point of the nasal feature. The symmetry analysis subunit is used to determine the first degree of symmetry difference of the nasal feature regions based on the distribution of feature points in the first and second regions. The distance determination subunit is used to determine the target movement distance of the camera based on the first degree of symmetry difference in the nasal feature region; The camera movement subunit is used to move the camera along the target direction by the target movement distance to obtain the camera after adjusting the shooting position; the target direction is the direction closer to the area with fewer feature points in the first region and the second region.

[0012] Furthermore, the present invention also proposes a symmetric analysis subunit for: A first edge line is constructed based on each feature point in the first region, and a second edge line is constructed based on each feature point in the second region. Obtain the edge line similarity between the first edge line and the second edge line, and obtain the difference in the number of feature points between the first region and the second region; The degree of first symmetry difference in nasal feature regions is determined by using edge line similarity and quantity difference.

[0013] Furthermore, the present invention also proposes a distance determination subunit for: Obtain the single-step movement distance of the camera, and the degree change value of the first symmetry difference corresponding to the single-step movement distance; The target movement distance of the camera is obtained by multiplying the ratio of the first degree of symmetry difference to the degree change value by the single-step movement distance.

[0014] Furthermore, the present invention also proposes a camera adjustment unit for: In response to the fact that each of the first facial features does not include the nose feature, the second degree of symmetry difference of the facial feature region corresponding to each of the first facial features is obtained; The overall symmetry difference of the target user's face is determined by utilizing the degree of second symmetry difference in each facial feature region; The camera's shooting position is adjusted based on the degree of overall symmetry difference.

[0015] Furthermore, the present invention also proposes that, prior to the target user's target verification method being a face verification method, the system further includes: The interface display module is used to display the target interface; the target interface includes each candidate verification method. The method determination module is used to determine the target verification method for the target user in response to the target user's selection of candidate verification methods in the target interface.

[0016] The present invention has the following beneficial effects: In the smart store management system for internet cafes provided in this invention, the image acquisition module acquires multiple consecutive frames of first face images via a camera when a target user selects a face verification method. The sequence construction module detects facial features in each first face image, constructs a facial feature missing sequence, and identifies which facial features are severely missing. The camera adjustment module adjusts the camera's shooting position based on the first facial features in the missing sequence, overcoming the limitations of a fixed position. The user control module uses the adjusted camera to acquire a second face image to control the target user's entry. Thus, by adjusting the camera's shooting position, higher-quality face images with more complete facial features can be acquired, effectively avoiding facial feature missingness caused by image quality issues, thereby significantly improving the accuracy of identity recognition. Attached Figure Description

[0017] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a schematic diagram of the structure of a smart store management and control system for internet cafes, provided as an embodiment of the present invention. Figure 2 This is a schematic diagram of the structure of a sequence construction module provided in one embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of a camera adjustment module provided in one embodiment of the present invention; Figure 4 This is a schematic diagram of the structure of a camera adjustment unit provided in one embodiment of the present invention. Detailed Implementation

[0019] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a smart store management system for internet cafes proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0021] In traditional smart access control systems, fixed-position cameras often fail to capture complete facial feature data due to changes in user posture or environmental interference. This manifests as persistent gaps in key facial feature regions across consecutive frame sequences. Such feature gaps directly impact the input data quality of facial recognition algorithms, leading to accumulated errors in feature vector extraction and consequently reducing the reliability of the identity verification process.

[0022] For example, in a fixed binocular camera system deployed at the entrance of an internet cafe, when a user stands at a 30-degree angle to the side, the outline of the nose bridge appears discontinuously broken in five consecutive frames, and the loss rate of the outer canthus feature point of the right eye reaches 60%. At this point, the 3D reconstruction model output by the facial feature detection algorithm shows a significant difference in the completeness of the left half of the face data compared to the right half, leading to a misjudgment of a fake face. Because the system cannot autonomously correct the acquisition angle, it can only passively process the issue by increasing the brightness of the supplementary lighting or providing voice prompts to the user to adjust their position, which prolongs the verification time.

[0023] To address the aforementioned issues, this invention first analyzes the data acquisition shortcomings of traditional fixed camera systems in dynamic scenes, finding that passive compensation mechanisms are insufficient to effectively handle feature loss caused by changes in user posture. Considering the possibility of dynamic adjustment of hardware parameters, this invention explores establishing a dynamic evaluation model for facial feature loss through multi-frame image analysis, thereby guiding the active adjustment of the camera's physical position. Compared to passive solutions that simply increase supplementary lighting intensity or rely on voice prompts, this invention chooses to construct a facial feature loss sequence as a control signal, optimizing the camera's viewing angle through a real-time feedback mechanism to ensure the integrity of feature acquisition. This scheme identifies key facial regions requiring priority compensation through facial feature loss detection in consecutive multi-frame images, and generates camera movement commands accordingly, forming a closed-loop control system.

[0024] To address this issue, the present invention provides a smart store management and control system for internet cafes. For example... Figure 1 As shown, the smart store management system 100 for internet cafes includes: The image acquisition module 110 is used to acquire multiple consecutive frames of the first face image of the target user through a camera, in response to the target user's target verification method being a face verification method.

[0025] In this module, the target user refers to an individual who plans to enter an internet cafe and needs to undergo identity verification to determine whether they are allowed to enter.

[0026] The target verification method refers to the method that a user chooses to verify their identity when entering an internet cafe, which may include facial recognition verification, ID card verification, etc.

[0027] The first face image in a series of frames refers to the image acquisition module 110 controlling the camera to continuously capture multiple face images of the user within a certain period of time after the target user selects the face verification method. These images serve as the basic data for subsequent analysis.

[0028] Specifically, when a target user arrives at the entrance of an internet cafe and prepares to enter, the system will prompt the target user to select a verification method. If the target user selects the facial verification method, the image acquisition module 110 will send a command to the camera to continuously capture multiple facial images of the target user over a period of time. These facial images constitute the first facial image of the target user in multiple consecutive frames, serving as the basic data for subsequent analysis and processing.

[0029] This method utilizes the continuous shooting function of the camera, and through software programming, sets the shooting time interval and frame rate, enabling the camera to acquire multiple frames of facial images according to preset requirements. This allows for the capture of the target user's facial features under different postures and expressions, providing a rich data sample for subsequent accurate analysis of missing facial features.

[0030] The sequence construction module 120 is used to detect facial features in each first face image and construct a facial feature missing sequence for the target user; the facial feature missing sequence includes first facial features that are severely missing in multiple consecutive frames of first face images.

[0031] In this module, facial feature detection refers to using specific algorithms to analyze each frame of the first face image and identify the facial features contained in the image, such as the shape and position of key parts like the eyes, nose, and mouth.

[0032] A facial feature missing sequence refers to a sequence formed by recording severely missing first facial features in multiple consecutive frames of first face images based on the detection results of facial features, and arranging them in a certain order. It is used to reflect the missing facial features of a target user in different frames of first face images.

[0033] The first facial feature refers to the facial feature that is marked as severely missing in the facial feature missing sequence, and is the key element in constructing the facial feature missing sequence.

[0034] Specifically, the sequence construction module 120 receives multiple consecutive frames of first face images from the image acquisition module 110, and uses a specific facial feature detection algorithm to analyze each frame of the first face image, identifying the shape and position of key parts such as the eyes, nose, and mouth. Then, based on the detection results, it records the first facial features that are severely missing in the multiple consecutive frames of the first face images, and arranges them in a certain order (such as the order of shooting time) to construct a sequence of missing facial features.

[0035] Among them, feature extraction algorithms based on computer vision and image processing technologies are used to extract key information of facial features through pixel analysis of face images. By setting a threshold for judging feature loss, when the completeness of a certain facial feature in multiple frames of the first face image is lower than the threshold, it is judged as severely missing and recorded in the facial feature loss sequence, thereby reflecting the missing facial features of the target user in different frames of the first face image.

[0036] The camera adjustment module 130 is used to adjust the shooting position of the camera based on each first facial feature in the missing facial feature sequence.

[0037] In this module, the shooting position includes parameters such as the camera's angle, height, and distance in space. These parameters determine the angle and range of the image captured by the camera.

[0038] Specifically, the camera adjustment module 130 acquires the facial feature missing sequence generated by the sequence construction module 120 and analyzes the reasons for the missing first facial features in the sequence. For example, if the eye features are severely missing, it may be inferred that the camera angle is too high or too low. Based on the analysis results, the camera's angle, height, distance, and other parameters in space are adjusted to change the camera's shooting position, thereby improving the shooting angle and range and acquiring a more complete facial image.

[0039] This involves basing the analysis on the geometric relationship between facial features and camera shooting position, as well as the imaging principle. Different facial feature deficiencies correspond to different shooting position issues. By establishing a mapping relationship between feature deficiencies and shooting position adjustments, mechanical devices or electronic control methods can be used to change camera parameters, thereby achieving precise adjustment of the shooting position.

[0040] User control module 140 is used to control the entry of the target user by using the second facial image of the target user obtained by the camera after adjusting the shooting position.

[0041] In this module, the second face image refers to the face image of the target user that is acquired again after the camera adjusts its shooting position. Compared with the first face image, it has higher quality and more complete facial features.

[0042] Specifically, after the camera adjustment module 130 completes the shooting position adjustment, the user control module 140 triggers the camera again to acquire a second facial image of the target user. Then, it performs identity recognition and verification on this second facial image, comparing the recognition result with the system's stored legitimate user information (e.g., information on all adults who have not committed crimes). If the comparison is successful, the target user is allowed to enter the internet cafe; if the comparison fails, the target user is denied entry.

[0043] This method utilizes high-quality second-person face images and employs facial recognition algorithms (such as feature point matching algorithms and deep learning algorithms) to extract facial feature information from these images. This information is then compared with pre-stored facial feature templates of legitimate users in a database. Based on the similarity threshold, the method determines whether the target user's identity is legitimate, thereby controlling the target user's entry into the internet cafe.

[0044] This invention analyzes the missing sequences of facial features in consecutive multi-frame facial images and dynamically adjusts the camera position to optimize the quality of facial image acquisition, thereby solving the problem of missing facial features caused by user posture or environmental factors in fixed cameras and improving the accuracy of identity verification.

[0045] As an example, an adjustable-angle high-definition camera is installed at the entrance of an internet cafe, and the camera is connected to a smart store management system. When a user selects facial verification, the system activates the camera and continuously captures the user's facial images at a rate of 3 frames per second, obtaining 15 frames of the first facial image within 5 seconds.

[0046] Next, facial feature detection is performed on each frame of the first face image, extracting feature points from key areas such as the eyes, nose, mouth, and chin. The number and distribution of feature points in each feature area are calculated and compared with a standard facial model to determine the feature completeness of each feature area. Based on the detection results of 15 consecutive frames, a facial feature missing sequence is constructed. For example, if it is found that the number of feature points in the user's right eye area is less than 80% of that in more than 50% of the frames, the right eye area is marked as a severely missing first facial feature and recorded in the facial feature missing sequence.

[0047] The facial feature loss sequence is then analyzed to determine the camera adjustment strategy. For example, if the right eye area is found in the facial feature loss sequence, the angle at which the camera needs to be adjusted to the left or right is calculated. The adjustment range is dynamically calculated based on the degree of loss; for example, for every 10% loss rate detected, the camera is adjusted 1 degree to the left or right.

[0048] Finally, after the camera adjusts its position, it captures another image of the user's face as a second facial image. Feature extraction and quality assessment are performed on this second facial image to ensure it contains complete facial feature information. Finally, based on the high-quality second facial image, identity verification is conducted to determine whether the user is allowed to enter the internet cafe.

[0049] In this embodiment, when a target user selects a face verification method, the image acquisition module 110 acquires multiple consecutive frames of first face images via a camera; the sequence construction module 120 detects facial features in each first face image, constructs a facial feature missing sequence, and identifies which facial features are severely missing; the camera adjustment module 130 adjusts the camera's shooting position based on the first facial features in the facial feature missing sequence, overcoming the limitations of a fixed position; and the user control module 140 uses the adjusted camera to acquire a second face image to control the target user's entry. Thus, by adjusting the camera's shooting position, higher-quality face images with more complete facial features can be acquired, effectively avoiding facial feature missingness caused by image quality issues, thereby significantly improving the accuracy of identity recognition.

[0050] In some of the above-described solutions of the present invention, when constructing a facial feature missing sequence through multiple consecutive frames of first face images, if only the missing facial features appearing in each frame are simply counted, misjudgment is likely to occur due to the randomness of a single frame image, making it impossible to accurately screen out facial features that are continuously and severely missing, thus affecting the accuracy of subsequent camera adjustments.

[0051] In this regard, such as Figure 2 As shown, the present invention further proposes a sequence construction module 120, comprising: The feature detection unit 121 is used to detect facial features in each first face image and determine the degree of feature loss of each facial feature in each first face image. The probability determination unit 122 is used to determine the probability of severe missing facial features based on the degree of feature loss of facial features in each first face image. Sequence construction unit 123 is used to sort each first facial feature with a severe missing probability greater than a first preset probability threshold according to the order of severe missing probability to obtain the facial feature missing sequence of the target user.

[0052] In this embodiment, the degree of feature loss refers to the extent to which various facial features (such as eyes, nose, mouth, etc.) are missing in each frame of the first face image compared to their complete and clear state. It is used to measure the completeness of facial features in the first face image. For example, a specific facial feature detection algorithm can be used to analyze the first face image. The algorithm compares the detected facial features with a pre-defined complete facial feature model, and comprehensively determines the degree of feature loss for each facial feature by calculating various indicators such as the number of missing key points (e.g., the corners of the eyes, the tip of the nose, etc.) and the degree of positional deviation.

[0053] The probability of severe missing features refers to the likelihood that a certain facial feature is severely missing in multiple consecutive frames of first-face images. It reflects the frequency and severity of the missing features in the overall first-face image. For example, for each facial feature, the degree of feature missing is statistically analyzed across all first-face images. Based on a pre-defined severe missing criteria (e.g., the feature missing degree reaches a certain value), the proportion of times the facial feature meets the severe missing criteria in each frame of the first-face image is calculated out of the total number of frames. This proportion is the probability of severe missing features for that facial feature. For example, if the eye feature is severely missing in 5 out of 10 frames of first-face images, then the probability of severe missing features for the eye feature is 50%.

[0054] The first preset probability threshold is a pre-set standard value for screening facial features. Only when the probability of severe missing facial features is greater than the first preset probability threshold will the facial feature be included in the facial feature missing sequence.

[0055] As an example, we first perform facial feature detection on each first face image to determine the degree of feature loss for each facial feature in each first face image. Specifically, we can use deep learning algorithms to detect facial features in each frame of the first face image, identifying the position and shape of key facial features such as the eyes, nose, and mouth. Then, by comparing the detected facial features with a standard face model, we calculate the degree of loss for each facial feature.

[0056] Then, for each facial feature, the probability of severe missing features is determined based on the degree of feature loss in each first face image. For example, the degree of missing features of each facial feature can be statistically analyzed in multiple consecutive frames. If a feature has a high degree of missing features in most frames, then the feature is considered to have a high probability of severe missing features.

[0057] Finally, the first facial features with a severe missing probability greater than a first preset probability threshold are sorted according to their severe missing probability to obtain the facial feature missing sequence of the target user. Further, the first preset probability threshold can be set to 0.7, and facial features with a severe missing probability exceeding 0.7 can be included in the facial feature missing sequence and sorted from high to low according to their severe missing probability.

[0058] This embodiment accurately identifies severely missing facial features in face images and prioritizes these features. This provides clear guidance for subsequent camera adjustments, allowing for targeted improvements in face image quality and enhancing the accuracy of face recognition. Furthermore, by setting a first preset probability threshold, slightly missing facial features can be filtered out, focusing on more impactful issues and improving system efficiency.

[0059] In some of the solutions described above in this invention, when determining the degree of feature loss of facial features in each first face image, there may be a problem that relying on only a single indicator may lead to an incomplete and inaccurate evaluation, which in turn affects the accuracy of subsequent calculation of the probability of severe feature loss and camera adjustment.

[0060] In response, the present invention further proposes a feature detection unit 121, used for: Obtain the current number of features of the second facial feature, the area of ​​the first feature of the second facial feature, and the area of ​​the second feature of the reference facial feature in the target first face image; the target first face image is any first face image, the second facial feature is any facial feature, and the reference facial features are all facial features except the second facial feature; The degree of feature loss of the second facial features in the target first face image is determined by using the ratio between the current number of features of the second facial features and the standard number of features of the second facial features, as well as the average difference between the area of ​​the first feature and the area of ​​each second feature.

[0061] In this embodiment, the ratio of the current number of features to the standard number of features reflects the proportion of missing features, and the difference between the first feature area and the reference feature area reflects the area deviation of the feature region. For example, when the second facial feature is an eye feature, the standard number of features can be preset to two feature points corresponding to both eyes. If the current number of features detects only one feature point, the ratio is 0.5. The first feature area can refer to the actual detected area of ​​the eye region, and the second feature area of ​​the reference facial feature can refer to the area of ​​the nose and mouth, etc., in the same face image. By combining the ratio of the number of features with the difference in the area, the severity of feature loss can be comprehensively assessed.

[0062] Furthermore, if there is only one standard feature for the second facial feature (e.g., mouth feature, nose feature), and no second facial feature is detected, that is, the current feature quantity of the second facial feature is 0, then the corresponding feature missing degree is directly marked as 1.

[0063] Specifically, when detecting the facial features of a target user, the current number of features of the second facial features in the first face image of the target is first obtained, and the ratio between the current number of features and the standard number of features is determined. For example, when the second facial feature is an eye feature, the standard number of features is preset to two feature points corresponding to both eyes. If the current number of features detects only one feature point, the ratio is 0.5. Simultaneously, the area of ​​the first feature of the second facial feature is calculated; for example, the area of ​​the left eye region is 50 pixels. The area of ​​the second feature of the reference facial features, such as the area of ​​the nose region being 150 pixels and the area of ​​the mouth region being 100 pixels, has a mean difference of 75 pixels. Further, based on the ratio of the number of features and the difference in area, the degree of feature loss is obtained. This allows for a more comprehensive quantification of feature loss, avoiding misjudgments caused by a single indicator, thereby improving the accuracy of subsequent camera position adjustments and ensuring the integrity of the facial features in the second face image.

[0064] The degree of feature loss in the second facial feature can be determined using the following formula 1: Formula 1 In formula 1, This is used to characterize the degree of feature loss of the r-th facial feature in the m-th first face image. The number of current features used to characterize the r-th facial feature in the m-th first face image. The number of standard features used to characterize the r-th facial feature The average difference between the area of ​​the first feature of the r-th facial feature in the m-th first face image and the area of ​​the second feature of each reference facial feature. Used to characterize an inverse proportional exponential function with base e. This represents the normalization function.

[0065] in, The smaller the value, the more incomplete the r-th facial feature is in the m-th first face image, that is, the greater the degree of feature loss. The larger the value, the greater the area difference between the r-th facial feature and other facial features in the m-th first face image, that is, the greater the degree of feature loss.

[0066] As an example, suppose the second facial feature is the eyes, the current feature count is the number of eyes detected, and the standard feature count is the preset number of eyes. Calculate the ratio of the current feature count to the standard feature count to obtain the feature count ratio.

[0067] The first feature area is the detected eye region area, and the second feature area is the area of ​​other facial features (such as the nose, mouth, etc.). The mean difference between the first feature area and each of the second feature areas is calculated to obtain the mean area difference.

[0068] Therefore, by combining the feature quantity ratio and the mean area difference, the degree of feature loss of the eye features can be obtained using Formula 1 above. The higher the degree of feature loss, the lower the integrity of the eye features in the image.

[0069] This embodiment accurately assesses the degree of missing facial features in a face image, providing a basis for adjusting the camera's shooting position and improving the accuracy of face recognition. By combining two dimensions—feature quantity and feature area—the integrity of facial features is comprehensively evaluated, avoiding misjudgments that may result from a single indicator. This method is applicable to various facial features, flexibly addressing different face recognition scenarios and enhancing the system's adaptability and robustness.

[0070] In some of the solutions described above in this invention, when determining the probability of severe loss of facial features, if only the average value of the degree of feature loss is used, it may not be able to effectively capture the dynamic trend of feature loss, resulting in insufficient accuracy of probability calculation and thus affecting the timeliness of camera adjustment.

[0071] In response, the present invention further proposes a probability determination unit 122, used for: The average value of the feature loss degree of the second facial feature in each first face image is obtained by averaging the feature loss degree of the second facial feature; the second facial feature is any one facial feature. The feature loss degree of the second facial features in each first face image is fitted in chronological order to obtain a linear curve of the feature loss degree of the second facial features. The probability of severe loss of the second facial feature is determined by using the average value of the feature loss degree and the slope of the feature loss degree line.

[0072] In this embodiment, the mean processing reflects the overall level of feature loss by calculating the arithmetic mean of the degree of loss of the same facial feature across multiple frames of images. The fitting process employs linear regression, transforming the feature loss data over time into a linear equation, with the slope representing the dynamic trend of the loss degree. When calculating the probability of severe loss using a combination of slope and mean, weighting coefficients can be set, and the product or weighted sum of the two can be used as the probability value. For example, when the slope is positive and the mean exceeds a preset threshold, the probability value increases significantly.

[0073] Specifically, after acquiring multiple frames of the first face image, the degree of feature loss of the second facial features in each frame is calculated. By averaging, the average degree of loss of the second facial feature across multiple consecutive frames is obtained, reflecting its overall loss status. Furthermore, the feature loss degrees of each frame are arranged chronologically and fitted to a linear curve representing the degree of feature loss, with the slope reflecting the rate of change of the loss degree over time. A positive slope indicates an increasing trend in the degree of loss; a negative slope indicates a gradually decreasing degree of loss. Combining the average value and the slope allows for a comprehensive assessment of the severity of feature loss. This method improves the accuracy of probability calculation through a dual assessment of dynamic trends and static levels, ensuring that the camera adjusts in a timely manner to acquire complete facial features.

[0074] Furthermore, the probability of severe loss of the second facial feature can be determined using the following formula 2: Formula 2 In formula 2, Used to characterize the probability of severe loss of the r-th facial feature. M is used to characterize the degree of feature loss of the r-th facial feature in the m-th first face image, and M is used to characterize the number of first face images. The slope of the feature loss line obtained by fitting the feature loss degree of M first face images. Represents the normalization function. This represents an exponential function with base e.

[0075] in, The larger the value, the greater the degree of feature loss of the r-th facial feature in the M first face images, and therefore the greater the probability of severe feature loss of the second facial feature. The larger the value, the more the degree of feature loss of the r-th facial feature shows a gradually increasing trend, and therefore the greater the probability of severe feature loss of the second facial feature.

[0076] As an example, the mean value of the feature loss of the second facial feature in each first face image is first calculated to obtain the average value of the feature loss of the second facial feature. The second facial feature can be any facial feature. For example, the eyes can be selected as the second facial feature. The feature loss of the eyes in multiple consecutive frames of the first face image is calculated, and then the average value is obtained.

[0077] Next, the feature loss degree of the second facial features in each first face image is fitted in chronological order to obtain a straight line showing the feature loss degree of the second facial features. Specifically, the time point of each frame can be used as the horizontal axis, and the corresponding feature loss degree can be used as the vertical axis to perform linear fitting to obtain a straight line.

[0078] Finally, using the average degree of feature loss and the slope of the feature loss line, the probability of severe loss of the second facial feature is determined by Formula 2 above. This allows for a more accurate determination of severe loss by comprehensively considering the overall level and trend of feature loss.

[0079] This embodiment enables a more accurate assessment of the probability of severe facial feature loss. By analyzing multiple frames of images, the randomness that may exist in single-frame images is avoided. Furthermore, by combining the average level and trend of feature loss, the assessment results are more comprehensive and reliable. This helps to more accurately adjust the camera's shooting position and improve the accuracy of face recognition.

[0080] In some of the solutions described above in this invention, when adjusting the camera shooting position, the existing solutions cannot effectively combine the specific distribution of nasal features to achieve precise adjustment, resulting in insufficient camera position adjustment accuracy and affecting the quality of subsequent face image acquisition.

[0081] In this regard, such as Figure 3 As shown, the present invention further proposes a camera adjustment module 130, including: The probability comparison unit 131 is used to compare the probability of severe loss of the nose feature with a second preset probability threshold in response to each first facial feature including the nose feature, and to obtain the probability comparison result. The voice prompt unit 132 is used to initiate a voice prompt to the target user in response to a probability comparison result indicating that the probability of severe loss of nasal features is greater than a second preset probability threshold, so as to instruct the target user to look directly at the camera. The camera adjustment unit 133 is used to adjust the shooting position of the camera based on the feature point distribution of the nasal features in response to the probability comparison result indicating that the probability of severe loss of nasal features is not greater than a second preset probability threshold.

[0082] In this embodiment, since the facial area is considered as a whole, all facial features will shift when the face is turned to the side. For example, when the face shifts significantly to the side, features in the eye area, nose area, and mouth area will also shift accordingly. Furthermore, with a slight turn, the eyes begin to be partially obscured, while the mouth area remains largely unaffected. However, with moderate and severe turns, the degree of eye obscuration increases, and the eyes may become completely obscured.

[0083] The nose area is less likely to be completely obscured. Due to its stable bony structure and strong resistance to obscuration, the nose area retains almost all its core features, experiencing only slight perspective compression, making it one of the most reliable facial feature areas. Therefore, the camera should be adjusted accordingly based on whether the nose is included in the primary facial features.

[0084] The core function of the probability comparison unit 131 is to determine the probability of severe loss of nasal features. After the system identifies nasal features from the first facial features, it calls a preset second probability threshold. This second preset probability threshold can be configured according to the facial recognition accuracy requirements of the actual internet cafe scenario, for example, set to 70% to balance recognition accuracy and user experience. The real-time calculated probability of severe loss of nasal features is then compared with this second preset probability threshold. When the probability comparison result indicates that the probability of severe loss of nasal features is greater than the second preset probability threshold, the voice prompt unit 132 will trigger the preset voice command module and play prompt content to the target user through the audio equipment in the corresponding area of ​​the internet cafe, such as "Please look directly at the camera to complete the identity verification", guiding the user to adjust the orientation of their face to ensure that the nasal features can be clearly captured.

[0085] When the probability comparison result shows that the probability of severe missingness does not exceed the second preset probability threshold, the camera adjustment unit 133 will calculate the deviation value between the current camera shooting angle and the user's face based on the feature point distribution data of the nose features, such as the key coordinate information of the two ends of the nose wings and the tip of the nose. Then, it will drive the camera to make fine adjustments to the angle in the horizontal or vertical direction until a complete and clear facial image of the nose features is obtained, so as to provide reliable image data support for subsequent user identity management.

[0086] As an example, when adjusting the camera's shooting position based on each first facial feature in the missing facial feature sequence, the system first determines whether each first facial feature includes a nasal feature. If a nasal feature is included, the probability of severe missing nasal features is compared with a second preset probability threshold to obtain a probability comparison result.

[0087] Furthermore, when the probability comparison result indicates that the probability of severe loss of nasal features is greater than a second preset probability threshold, a voice prompt is issued to the target user to instruct the user to look directly at the camera. For example, by broadcasting the prompt message "Please look directly at the camera to complete identity verification," the user is guided to adjust their facial orientation.

[0088] When the probability comparison result indicates that the probability of severe loss of nasal features is not greater than a second preset probability threshold, the camera's shooting position is adjusted based on the distribution of nasal feature points. Specifically, the camera's movement direction and distance can be determined by analyzing the symmetry of the nasal feature point distribution to obtain a better shooting angle.

[0089] Therefore, the solution of this invention can take different processing measures according to the degree of loss of nasal features. When nasal features are severely lost, the user is guided to adjust their posture through voice prompts; when the loss of nasal features is relatively minor, the shooting effect is optimized by adjusting the camera position. This flexible processing method can effectively improve the quality of facial image acquisition.

[0090] This embodiment enables the implementation of corresponding processing measures based on the degree of missing nasal features, effectively improving the quality of facial image acquisition. When nasal features are severely missing, voice prompts guide the user to adjust their posture, avoiding recognition failures caused by improper user posture. When the degree of missing nasal features is relatively mild, the camera position is adjusted to optimize the shooting effect and improve the accuracy of facial recognition. This flexible processing method can adapt to different recognition scenarios, improving the system's recognition efficiency and user experience.

[0091] In some of the solutions described above in this invention, when adjusting the camera's shooting position based on the distribution of feature points of the nose, the uneven distribution of feature points or inaccurate region division may affect the accuracy of the camera's moving distance, making it impossible to effectively adjust the shooting position to obtain a complete facial feature image.

[0092] In this regard, such as Figure 4 As shown, the present invention further proposes a camera adjustment unit 133, comprising: The region division subunit 1331 is used to divide the nasal feature region corresponding to the nasal feature into a first region to the left of the central feature point and a second region to the right of the central feature point, based on the central feature point of the nasal feature. The symmetry analysis subunit 1332 is used to determine the first degree of symmetry difference of the nasal feature region based on the distribution of feature points in the first region and the second region. Distance determination subunit 1333 is used to determine the target movement distance of the camera based on the first degree of symmetry difference of the nasal feature region; The camera movement subunit 1334 is used to move the camera along the target direction by a target movement distance to obtain the camera after adjusting the shooting position; the target direction is the direction closer to the area with fewer feature points in the first region and the second region.

[0093] In this embodiment, the region segmentation subunit 1331 uses the central feature point of the nose (such as the tip of the nose) as a reference. In the two-dimensional facial image, an image coordinate segmentation algorithm generates a virtual vertical segmentation line passing through the central feature point, dividing the nose feature region into two parts: a first region to the left of the central feature point and a second region to the right. The boundary between the two parts is a vertical virtual straight line passing through the central feature point. The symmetry analysis subunit 1332 calculates the first degree of symmetry difference by statistically analyzing the number of feature points, the distribution density of feature points, and the positional deviation of key feature points (such as the alar and columella) between the first and second regions. For example, the difference in the number of feature points between the two regions can be used as a quantification index of the difference. The distance determination subunit 1333 uses a preset mapping table between the degree of symmetry difference and the moving distance. The greater the first degree of symmetry difference, the greater the corresponding target moving distance, ensuring that the camera can quickly adjust to a symmetrical shooting position.

[0094] Specifically, after acquiring the feature point distribution data of the nose features, the region segmentation subunit 1331 first locates the tip of the nose as the central feature point. In the two-dimensional facial image, based on the central feature point, a virtual vertical segmentation line is generated along the vertical direction of the image. This vertical segmentation line passes through the central feature point and is perpendicular to the horizontal direction of the image, thus dividing the nose feature region into a first region on the left and a second region on the right in the two-dimensional plane. The left side of the segmentation line is the first region, and the right side is the second region. The symmetry analysis subunit 1332 counts the number of feature points in the first and second regions respectively. If the first region has 15 feature points and the second region has 10 feature points, the ratio of the difference between the two (5) to the total number of feature points (25), which is 20%, is taken as the first degree of symmetry difference. The distance determination subunit 1333 determines the target movement distance to be 4 millimeters according to a preset rule, such as an increase of 2 millimeters in the movement distance for every 10% increase in the degree of symmetry difference. The camera moving subunit 1334 determines that the number of feature points in the second region is less through image analysis, and determines the target direction to be closer to the second region. Then, it drives the horizontal moving mechanism of the camera to move 4 mm in this direction to complete the shooting position adjustment, and finally achieves symmetrical presentation of the nose feature area in the camera image, improving the stability of subsequent feature recognition.

[0095] As an example, using the central feature point of the nose as a benchmark, the nasal feature region corresponding to the central feature point is divided into a first region to the left of the central feature point and a second region to the right of the central feature point. Further, based on the feature point distribution between the first and second regions, the degree of first symmetry difference of the nasal feature regions is determined. Specifically, the degree of first symmetry difference can be quantified by calculating the difference in the number of feature points and the difference in feature point distribution density between the first and second regions.

[0096] Therefore, the target movement distance of the camera is determined based on the degree of symmetry difference in the nasal feature region. For example, a mapping table between the degree of symmetry difference and the movement distance can be preset, and then the corresponding target movement distance can be obtained from the mapping table according to the magnitude of the first degree of symmetry difference.

[0097] Finally, the camera is moved a target distance along the target direction to obtain the adjusted camera position. The target direction is the direction closer to the area with fewer feature points in the first and second regions. This method allows the camera to be more precisely positioned to capture the center of the user's face, resulting in a clearer and more complete facial image.

[0098] This embodiment enables the camera's shooting position to be automatically adjusted based on the actual distribution of the user's facial features. This dynamic adjustment mechanism effectively addresses the problem of misalignment between the user's face and the camera, improving the quality of facial image acquisition. Consequently, it enhances the accuracy and reliability of facial recognition, improves the user experience, and also increases the management efficiency of the internet cafe access control system.

[0099] In some of the solutions described above in this invention, the nasal feature region is divided into a first region and a second region, and the degree of symmetry difference is determined based on the distribution of feature points, thereby adjusting the camera position. However, in this process, relying solely on the number of feature points or simple positional relationships may not fully reflect the symmetry differences of the nasal feature region, resulting in insufficient accuracy in calculating the degree of symmetry difference and affecting the accuracy of camera position adjustment.

[0100] In response, the present invention further proposes a symmetry analysis subunit 1332, used for: A first edge line is constructed based on each feature point in the first region, and a second edge line is constructed based on each feature point in the second region. Obtain the edge line similarity between the first edge line and the second edge line, and obtain the difference in the number of feature points between the first region and the second region; The degree of first symmetry difference in nasal feature regions is determined by using edge line similarity and quantity difference.

[0101] In this embodiment, the first edge line and the second edge line are generated by connecting feature points within their corresponding regions. The edge line similarity can be obtained by calculating the shape matching degree between the two edge lines, for example, by using the Hausdorff distance algorithm to quantify the overall difference between the first edge line and the second edge line. The difference in number is calculated by taking the absolute value after counting the number of feature points in the first region and the second region.

[0102] Specifically, when constructing the first and second edge lines, feature point coordinates are sequentially connected to form a continuous curve, capturing the contour differences between the left and right sides of the nose. The edge line similarity calculation is achieved through the following steps: 1. For each feature point on the first edge line, calculate its Euclidean distance to all feature points on the second edge line, and take the minimum value (i.e., the shortest distance from a single point to the opposite edge line); 2. Calculate the arithmetic mean of the shortest distances corresponding to all points on the first edge line to obtain the overall average minimum distance between the first and second edge lines; 3. Calculate the overall average minimum distance between the second and first edge lines symmetrically; 4. Take the maximum value of the two overall average minimum distances as the overall difference measure between the two edge lines, and normalize its reciprocal (e.g., map it to the [0,1] interval) to obtain the edge line similarity between the first and second edge lines. The quantity difference reflects the difference in the density of feature point distribution; for example, sparse feature points on the left side may lead to an incomplete edge line. Combining edge line similarity with the quantity difference, the first symmetry difference degree simultaneously covers shape differences and distribution differences, providing a more comprehensive assessment of nasal symmetry. For example, when the curvature of the left edge line is high and the number of feature points is small, the degree of first symmetry difference increases, triggering the camera to move a further distance to the left, ensuring that the adjusted camera can capture complete nasal features and avoid recognition failure due to local missing features.

[0103] The degree of first symmetry difference in the nasal feature area can be determined using the following formula 3: Formula 3 In formula 3, The degree of first symmetry difference used to characterize the nasal feature region The number of feature points in the first region used to characterize the nasal feature area. The number of feature points in the second region used to characterize the nasal feature area. The first edge line in the first region used to characterize the nasal feature area The second edge line in the second region, used to characterize the nasal feature area, is ESS, which characterizes the edge line similarity calculated based on the above edge line similarity calculation process. This represents the normalization function.

[0104] The greater the similarity between the first and second edge lines, the smaller the degree of first symmetry difference in the nasal feature region; the greater the difference in the number of feature points between the first region and the second region, the greater the degree of first symmetry difference in the nasal feature region.

[0105] As an example, a first edge line is constructed based on the feature points in the first region, and a second edge line is constructed based on the feature points in the second region. Specifically, a B-spline curve fitting algorithm can be used to fit the feature points in the first and second regions to obtain smooth first and second edge lines.

[0106] Next, the edge line similarity between the first and second edge lines is obtained, and the difference in the number of feature points in the first region and the second region is also calculated. The edge line similarity can be measured using the aforementioned edge line similarity calculation process. The difference in the number of feature points is obtained by directly subtracting the number of feature points in the two regions.

[0107] Finally, using edge line similarity and quantity difference, the first degree of symmetry difference of the nasal feature region is determined by the above formula 3.

[0108] This embodiment enables more accurate quantification of the symmetry of the nasal feature region, providing a reliable basis for subsequent adjustments to the camera's shooting position. This improves the accuracy of facial recognition, reduces recognition failures due to missing facial features, and ultimately enhances the overall performance and user experience of the internet cafe smart store management system.

[0109] In some of the solutions described above in this invention, the target moving distance of the camera is adjusted based on the first degree of symmetry difference of the nasal feature region. However, in this process, how to dynamically determine the appropriate moving distance according to the change of the degree of symmetry difference, and avoid the problem of low adjustment efficiency or poor effect caused by the moving distance being too large or too small, has not yet been effectively solved.

[0110] In response, the present invention further proposes a distance determination subunit 1333, used for: Obtain the single-step movement distance of the camera, and the degree change value of the first symmetry difference corresponding to the single-step movement distance; The target movement distance of the camera is obtained by multiplying the ratio of the first degree of symmetry difference to the degree change value by the single-step movement distance.

[0111] In this embodiment, the single-step movement distance is the minimum unit distance that the camera moves each time, which can be preset through experiments; the degree change value represents the expected change in the degree of first symmetry difference when the camera moves a single step distance, which can be obtained by fitting historical data or testing; by calculating the ratio of the degree of first symmetry difference to the degree change value, the number of single-step movements required for the current degree of symmetry difference can be quantified, and then the number of single-step movements can be multiplied by the single-step movement distance to obtain the final target movement distance.

[0112] Specifically, when adjusting the camera position, the first step is experimentally determined to correlate the single-step movement distance with the corresponding degree of change. For example, a single-step movement distance of 1 cm reduces the first degree of symmetry difference by 0.2. When the first degree of symmetry difference in the nasal feature region is detected to be 0.6, the ratio (0.6 / 0.2=3) determines that 3 single-step movements are needed, with a final target movement distance of 3 cm. Thus, the camera can dynamically adjust its movement distance based on the current degree of symmetry difference, avoiding multiple adjustments due to insufficient movement distance or overshoot due to excessive movement distance, thereby improving adjustment accuracy and efficiency. This scheme quantifies the relationship between the degree of difference and the movement distance, ensuring the camera moves with the optimal step size to quickly reach the position where symmetry difference is minimized, thereby improving facial image quality and verification accuracy.

[0113] The target movement distance of the camera can be determined using the following formula 4: Formula 4 In formula 4, Used to characterize the target movement distance of the camera. The degree of first symmetry difference is used to characterize the nasal feature region, and G is used to characterize the degree change of the first symmetry difference corresponding to a single step movement distance. Used to characterize the distance traveled in a single step.

[0114] As an example, when determining the target movement distance of the camera, the single-step movement distance is first preset to a fixed value, which is set to 0.5 cm based on the accuracy range of the camera's mechanical structure. Experimental testing is used to obtain the degree of change in the first symmetry difference corresponding to the single-step movement distance. For example, when the camera moves 0.5 cm, the first symmetry difference degree can decrease by 0.2. In specific operation, when the first symmetry difference degree of the detected nasal feature region is 0.8, the target movement distance is calculated as 0.8 divided by 0.2 and then multiplied by 0.5 cm, resulting in a movement of 2 cm. At this point, the stepper motor of the camera bracket is controlled to move 2 cm along the target direction, bringing the camera closer to the side with sparse nasal feature points, thereby optimizing the shooting angle.

[0115] This embodiment enables dynamic adjustment of camera movement based on the symmetry differences in the nasal feature region, effectively solving the problem of missing facial features caused by shooting angle deviations. By quantifying the correlation between the degree of difference and the movement distance, the camera can accurately compensate for shooting angle deviations with each movement, thereby improving the integrity of nasal features in facial images, avoiding secondary image distortion caused by excessive or insufficient mechanical adjustment, and significantly improving the accuracy of identity verification.

[0116] In some of the solutions described above in this invention, a scheme for adjusting the camera movement distance based on the first degree of symmetry difference in the nasal feature region is proposed. However, no solution is proposed for how to adjust the camera shooting position when the first facial features do not include nasal features, resulting in poor accuracy in adjusting the camera shooting position.

[0117] In response, the present invention further proposes a camera adjustment unit 133, used for: In response to the fact that each of the first facial features does not include the nose feature, the second degree of symmetry difference of the facial feature region corresponding to each of the first facial features is obtained; The overall symmetry difference of the target user's face is determined by utilizing the degree of second symmetry difference in each facial feature region; The camera's shooting position is adjusted based on the degree of overall symmetry difference.

[0118] In this embodiment, when the first facial feature does not include a nose feature, the camera adjustment unit 133 first activates the facial feature region symmetry analysis mechanism. The second degree of symmetry difference is determined using the same calculation method as the first degree of symmetry difference in the above embodiment of the present invention, and will not be described in detail here. The comprehensive degree of symmetry difference is obtained by assigning corresponding weights to different facial feature regions (e.g., eye weight 40%, mouth weight 30%, cheek weight 30%), and then weighting and summing multiple second degrees of symmetry difference to form a comprehensive degree of symmetry difference that reflects the overall facial symmetry state. When adjusting the camera's shooting position, the length of movement is determined using the same calculation method as the target movement distance of the camera in the above embodiment of the present invention, and will not be described in detail here. The movement direction when adjusting the camera's shooting position is determined based on the direction that appears most frequently among the movement directions corresponding to each first facial feature. The movement direction corresponding to each first facial feature is determined using the same determination method as the target direction in the above embodiment of the present invention, and will not be described in detail here.

[0119] The degree of overall symmetry difference can be determined using the following formula 5: Formula 5 In formula 5, Used to characterize the degree of comprehensive symmetry difference The degree of second symmetry difference in the facial feature region corresponding to the nth first facial feature is used to characterize the value accumulated from 1 to N, and N is used to characterize the total number of first facial features.

[0120] in, for When calculating the second degree of symmetry difference, the weight coefficients are substituted into the calculation in order of increasing probability of severe loss of the first facial features; that is, the greater the probability of severe loss, the greater the corresponding weight coefficient.

[0121] As an example, when the first facial features do not include the nose feature, the eye and lip feature regions of the target user are automatically extracted, and the corresponding second symmetry difference coefficients are obtained respectively. Then, the second symmetry difference coefficients are weighted and summed according to preset weights to obtain the comprehensive symmetry difference value of the face. Based on the comprehensive symmetry difference value, the camera is controlled to rotate 5 degrees in the opposite direction of the target user's head tilt along the vertical axis, and translated 3 centimeters in the direction of the head deflection along the horizontal axis, so that the camera's optical axis is aligned with the midline of the user's face.

[0122] This embodiment effectively solves the problem of not being able to accurately adjust the camera position. By comprehensively analyzing the symmetry differences of multiple facial feature regions, it can intelligently correct the camera's shooting angle and orientation, ensuring that complete facial feature images can still be obtained under complex head postures, and significantly improving the environmental adaptability and recognition accuracy of the face verification system.

[0123] In some of the solutions described above in this invention, the smart store management system uses a fixed face verification method by default and does not provide an interactive mechanism for users to choose their own verification method. This results in users being unable to choose a suitable verification method based on their own conditions or preferences, thereby affecting verification efficiency and user experience.

[0124] In response, this invention further proposes that, prior to the face verification method, the target user's verification method includes: The interface display module is used to display the target interface; the target interface includes each candidate verification method. The method determination module is used to determine the target verification method for the target user in response to the target user's selection of candidate verification methods in the target interface.

[0125] In this embodiment, the target interface integrates multiple candidate verification methods through graphical interactive elements, including but not limited to facial recognition, fingerprint recognition, or ID card verification. The arrangement of candidate verification methods is generated based on preset priority rules, which are dynamically adjusted according to historical verification success rates. Selection is achieved through touch clicks or voice commands, and a control signal containing the identifier of the target verification method is generated after the operation is triggered.

[0126] Specifically, the target interface automatically activates when a user approaches the access control device. Candidate verification methods are arranged horizontally as icons, each labeled with its name. When a user selects the face verification method icon, the touch coordinates are converted into a selection command, which is then parsed to activate the camera. Further, the backend receives the target verification method identifier and invokes the corresponding verification algorithm resources; for example, when face verification is selected, the facial feature database is loaded. In this process, the determination of the target verification method precedes the camera image acquisition step, avoiding hardware initialization delays caused by verification method switching. For example, when the user selects fingerprint verification, the fingerprint sensor is directly invoked and the camera power is turned off, thereby reducing power consumption.

[0127] As an example, an interactive terminal device with an integrated touchscreen display is installed at the entrance of an internet cafe. When a user approaches the terminal, the graphics rendering engine is invoked to present candidate verification methods—including facial recognition, ID card verification, and fingerprint verification—as a proportionally scaled array of icons on the touchscreen. The verification method name is labeled below each icon, with the facial recognition icon featuring a dynamic halo effect to indicate the recommended option. When the user taps the facial recognition icon, the capacitive sensing layer of the touchscreen generates a contact coordinate signal. Based on the coordinate signal mapping, the target verification method is determined to be facial recognition.

[0128] This embodiment visualizes and makes the verification method selection process more interactive. By guiding users to actively select the appropriate verification method through a graphical interface, it avoids repeated operations caused by mismatched verification methods, reduces the invalid waiting time in the identity verification process, and enhances the recognition accuracy of user operation intentions through a dynamic feedback mechanism of interface elements, providing accurate verification method trigger signals for the subsequent face image acquisition stage.

[0129] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0130] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A smart store management and control system for internet cafes, characterized in that, The system comprises: An image acquisition module, configured to acquire, in response to a target user's target verification mode being a face verification mode, continuous multiple frames of first face images of the target user through a camera; A sequence construction module, configured to detect face features of each of the first face images respectively, and construct a face feature missing sequence of the target user; the face feature missing sequence includes first face features that are seriously missing in the continuous multiple frames of first face images; A camera adjustment module, configured to adjust a shooting position of the camera based on each of the first face features in the face feature missing sequence; A user control module, configured to control the target user to enter by acquiring second face images of the target user through the camera after the shooting position is adjusted. 2.The intelligent shop management and control system for internet cafes according to claim 1, characterized in that, The sequence construction module comprises: A feature detection unit, configured to detect face features of each of the first face images respectively, and determine a feature missing degree of each of the face features in each of the first face images; A probability determination unit, configured to determine, for each of the face features, a seriously missing probability of the face feature based on the feature missing degree of the face feature in each of the first face images; A sequence construction unit, configured to sort each of the first face features whose seriously missing probability is greater than a first preset probability threshold in a size order of the seriously missing probability, and obtain a face feature missing sequence of the target user. 3.The intelligent shop management and control system for internet cafes according to claim 2, characterized in that, The feature detection unit is configured to: acquire a current feature quantity of a second face feature, a first feature area of the second face feature, and a second feature area of a reference face feature in a target first face image; the target first face image is any one of the first face images, the second face feature is any one of the face features, and the reference face feature is each of the face features other than the second face feature; determine the feature missing degree of the second face feature in the target first face image by using a ratio between the current feature quantity of the second face feature and a standard feature quantity of the second face feature, and a mean value of differences between the first feature area and each of the second feature areas. 4.The intelligent shop management and control system for internet cafes according to claim 2, characterized in that, The probability determination unit is configured to: perform mean value processing on the feature missing degree of a second face feature in each of the first face images to obtain a feature missing degree average value of the second face feature; the second face feature is any one of the face features; fit the feature missing degree of the second face feature in each of the first face images in a time order to obtain a feature missing degree straight line of the second face feature; determine the seriously missing probability of the second face feature by using the feature missing degree average value and a slope of the feature missing degree straight line. 5.The intelligent shop management and control system for internet cafes according to claim 1, characterized in that, The camera adjustment module comprises: A probability comparison unit, configured to, in response to the first face features including a nose feature, compare a seriously missing probability of the nose feature with a second preset probability threshold to obtain a probability comparison result. The voice prompting unit is configured to initiate voice prompting to the target user to indicate that the target user looks at the camera in response to the probability comparison result indicating that the probability of severe absence of the nose feature is greater than the second preset probability threshold. The camera adjusting unit is configured to adjust the shooting position of the camera based on the distribution of feature points of the nose feature in response to the probability comparison result indicating that the probability of severe absence of the nose feature is not greater than the second preset probability threshold. 6.The intelligent shop management and control system for internet cafes according to claim 5, characterized in that, The camera adjusting unit includes: The region division subunit is configured to divide a nose feature region corresponding to the nose feature into a first region on the left of a center feature point of the nose feature and a second region on the right of the center feature point. The symmetry analysis subunit is configured to determine a first symmetry difference degree of the nose feature region based on the distribution of feature points of the first region and the second region. The distance determination subunit is configured to determine a target moving distance of the camera based on the first symmetry difference degree of the nose feature region. The camera moving subunit is configured to move the camera by the target moving distance in a target direction to obtain the camera after the shooting position is adjusted, wherein the target direction is a direction close to the region with less feature points between the first region and the second region.

7. The intelligent shop management and control system for internet cafes according to claim 6, characterized in that, The symmetry analysis subunit is configured to: construct a first edge line based on each feature point in the first region and construct a second edge line based on each feature point in the second region; obtain an edge line similarity between the first edge line and the second edge line and obtain a quantity difference value between the number of feature points in the first region and the number of feature points in the second region; and determine the first symmetry difference degree of the nose feature region by using the edge line similarity and the quantity difference value. 8.The intelligent shop management and control system for internet cafes according to claim 6, characterized in that, The distance determination subunit is configured to: obtain a single-step moving distance of the camera and a degree change value of the first symmetry difference degree corresponding to the single-step moving distance; and multiply the ratio of the first symmetry difference degree and the degree change value by the single-step moving distance to obtain the target moving distance of the camera. 9.The intelligent shop management and control system for internet cafes according to claim 5, characterized in that, The camera adjusting unit is configured to: obtain a second symmetry difference degree of each face feature region corresponding to each first face feature in response to the first face feature not including the nose feature; determine a comprehensive symmetry difference degree of the face of the target user by using the second symmetry difference degree of each face feature region; and adjust the shooting position of the camera based on the comprehensive symmetry difference degree.

10. The intelligent shop management and control system for internet cafes according to any one of claims 1-9, characterized in that, Before the target user's target verification mode is the face verification mode, the system further includes: an interface display module configured to display a target interface, wherein the target interface includes each candidate verification mode; a mode determination module configured to determine the target verification mode of the target user in response to a selection operation of the target user on the candidate verification mode in the target interface.

Images