An intelligent self-service correction terminal

Abstract

The present invention discloses an intelligent self-service correction terminal, which is equipped with an anti-counterfeiting detection device for face anti-fraud liveness detection, wherein the anti-counterfeiting detection device is configured to obtain a plurality of pictures to be verified collected by the camera device and pre-stored comparison pictures; identify human features in the pictures to be verified, wherein the human features include contour information, position information and human body part categories; create a cropping carrier based on the contour information and position information of the human features; use the cropping carrier to crop a plurality of pictures to be verified and the comparison pictures to obtain a plurality of environment verification pictures and environment comparison pictures; compare the plurality of environment verification pictures with the environment comparison pictures to calculate similarity; when the similarity of the plurality of environment verification pictures does not meet a threshold condition, output a detection result of forged face information; when the similarity of the plurality of environment verification pictures meets a threshold condition, output a detection result determined as liveness.

Description

Technical Field

[0001] The present invention belongs to the technical field of community correction equipment, and in particular relates to an intelligent self-service correction terminal. Background Art

[0002] The applicant has found that the existing self-service correction terminals have technical security loopholes, which can be used to forge records of registration and learning through abnormal channels, thus escaping management. However, the counties and grassroots offices responsible for the specific implementation of community correction activities face many practical difficulties, making it difficult to ensure the standardized implementation of community correction work. They rely too much on the self-consciousness of the correction subjects, lack of seriousness and coercive force, and easily cause leakage of management, thus making the corrections meaningless.

[0003] The technical security loopholes of existing self-service correction terminals are mainly technical defects in liveness detection. Self-service correction terminals use cameras to authenticate the current person online. However, there are more and more attacks on biometrics. The main goal of these attacks is to use counterfeit non-live input data such as pre-recorded videos and images to make liveness detection successful and complete the comparison of biometric information. Even if the person is not present, he can still complete the registration and learning, making the correction object out of management. Summary of the invention

[0004] The purpose of the present invention is to solve the technical security loopholes of existing self-service correction terminals, and to provide an intelligent self-service correction terminal by imitating non-living input data such as pre-recorded videos and images to successfully detect liveness and complete the comparison of biological information. The person himself can also complete the registration and learning without being present, thereby making the correction object out of management.

[0005] In order to solve the above problems, the present invention is implemented according to the following technical solutions:

[0006] The present invention provides an intelligent self-service correction terminal, comprising a terminal body, wherein the terminal body is provided with a camera device for liveness detection and recording the activities of a correction object; the terminal body is also provided with an anti-counterfeiting detection device connected to the camera device, and the anti-counterfeiting detection device is configured as follows:

[0007] Acquire multiple pictures to be verified captured by the camera device and pre-stored comparison pictures;

[0008] Identify human features in the image to be verified, where the human features include contour information, position information, and human body part categories;

[0009] Creating a cutting vector based on the contour information and position information of human body features;

[0010] Using a cropping carrier to crop multiple to-be-verified pictures and comparison pictures, a plurality of environment verification pictures and environment comparison pictures are obtained;

[0011] Compare the plurality of environment verification pictures with the environment comparison pictures, and calculate the similarity;

[0012] When the similarity of multiple environment verification images does not meet the threshold condition, the detection result of forged face information is output;

[0013] When the similarity of multiple environmental verification images meets the threshold condition, the output is a detection result that determines it to be a living body.

[0014] In a preferred implementation, the method of using a cropping carrier to crop multiple images to be verified and comparison images to obtain multiple environment verification images and environment comparison images specifically includes:

[0015] Adding a cropping carrier to the image to be verified, and setting the cropping carrier to be the content to be cropped on the image to be verified;

[0016] According to the position information of the cropping carrier, the image to be verified is cropped along the contour information in the cropping carrier, and a portion of the image to be verified within a range surrounded by the contour information of the cropping carrier is filtered to obtain an environment verification image;

[0017] Adding a cropping carrier to the comparison image, and setting the cropping carrier to be the content to be cropped on the comparison image;

[0018] The comparison picture is cropped along the contour information of the cropping carrier according to the position information of the cropping carrier, and the part of the comparison picture within the range surrounded by the contour information is filtered to obtain the environment comparison picture.

[0019] In a preferred implementation, comparing the plurality of environment verification pictures with the environment comparison pictures and calculating the similarity specifically includes:

[0020] Reduce the size of multiple images to be verified and compared to 64*64;

[0021] Convert the reduced image to be verified and the comparison image into grayscale images to obtain multiple grayscale images to be verified and comparison grayscale images;

[0022] Calculate the pixel averages of multiple grayscale images to be verified and the comparison grayscale images;

[0023] Traverse each pixel of multiple grayscale images to be verified and compared grayscale images, compare the pixel with the pixel average value, and obtain hash values ​​of multiple images to be verified and compared images;

[0024] The Hamming distances between the multiple to-be-verified images and the comparison images are calculated in sequence to obtain the similarities between the multiple to-be-verified images and the comparison images.

[0025] In a preferred implementation, the anti-counterfeiting detection device is configured to perform the following steps before performing the environmental comparison process:

[0026] Identify whether multiple images to be verified contain electronic screens;

[0027] If not, then the environmental comparison process is carried out;

[0028] If yes, identify the object features of the electronic screen, the object features including the outline of the electronic screen and the position of the electronic screen in the image to be verified;

[0029] Determining whether the contour information of the human body feature is located within the electronic screen contour of the object feature;

[0030] If yes, output the detection result of forged face information;

[0031] If not, the environment comparison process is performed.

[0032] In a preferred implementation, the intelligent self-service correction terminal further includes a liveness detection device, and the liveness detection device is configured as follows:

[0033] Receiving a certification request for a subject to be certified and corrected;

[0034] generating at least two random behaviors; each of the random behaviors indicates a facial action, and different random behaviors indicate different facial actions;

[0035] Randomly combining the at least two random behaviors to generate a random behavior sequence; sending the random behavior sequence to a display screen of the intelligent self-service correction terminal for display;

[0036] Acquire the video to be verified captured by the camera device; obtain the facial action sequence of the correction object using the video to be verified;

[0037] determining whether at least two facial actions included in the facial action sequence are the same as facial actions indicated by at least two random behaviors in the random behavior sequence, and;

[0038] Determining whether the order of at least two facial actions included in the facial action sequence in the facial action sequence is the same as the order of at least two random actions in the random action sequence in the random action sequence;

[0039] If they are the same, the authentication success result is output.

[0040] In a preferred implementation, the intelligent self-service correction terminal further includes a positioning device and a position monitoring device, wherein the positioning device is used to collect the current position information of the intelligent self-service correction terminal, and the position monitoring device is configured as follows:

[0041] Obtaining current location information of the positioning device;

[0042] Load the pre-stored geo-fence and determine whether the current location information is within the geo-fence;

[0043] If yes, a hit result is returned, and based on the hit result, the intelligent self-service correction terminal is instructed to perform a liveness detection process;

[0044] If not, a miss result is returned, and an abnormal position alarm message is generated and issued based on the miss result, and the external service of the intelligent self-service correction terminal is stopped.

[0045] In a preferred implementation, the intelligent self-service correction terminal further includes an identity information collection device and a data processing device, the data processing device is connected to the camera device, and the data processing device is configured to encrypt the video data of the camera device, including the following encryption method:

[0046] Acquire the video data currently collected by the camera device and the identity information of the correction object currently collected by the identity information collection device;

[0047] Encrypting the video data, and storing the encrypted video data locally in association with the identity information;

[0048] The data processing device is further configured to decrypt the encrypted video data, including the following decryption method:

[0049] When a data upload event is triggered, identify whether the video data stored locally and to be uploaded is encrypted;

[0050] If yes, decrypt the video data stored locally, and upload the decrypted video data and identity information based on the communication link;

[0051] If not, a video forgery alarm message is generated and issued.

[0052] In a preferred embodiment, the terminal body is provided with an image capture device for document scanning; the image capture device is pivotally connected to the terminal body to form a rotating pair; the image capture device can swing outward from the side wall of the terminal body around the pivot, and can also be stored and attached to the side wall of the terminal body.

[0053] In a preferred implementation, an intelligent self-service correction terminal also includes a document placement pad detachably connected to the terminal body; a document placement area mark is provided on the upper surface of the document placement pad; when the image capture device swings to the target position for document scanning, the image capture device correspondingly hovers above the document placement pad.

[0054] In a preferred implementation, the terminal body includes:

[0055] A base, on which a bracket body installed upright is provided;

[0056] A front display screen and a rear display screen for human-computer interaction, wherein the front display screen and the rear display screen are installed on the front and back of the bracket body respectively;

[0057] A camera device for photographing a human face, wherein the camera device is fixedly connected to the support body;

[0058] The image capture device is movably connected to the bracket body, and the image capture device is located between the front display screen and the rear display screen; an ID card recognition device and a fingerprint scanning device are integrated and installed on the base.

[0059] Compared with the prior art, the present invention has the following beneficial effects:

[0060] An intelligent self-service correction terminal of the present invention is equipped with an anti-counterfeiting detection device for face anti-fraud liveness detection, wherein the anti-counterfeiting detection device is configured to obtain multiple pictures to be verified collected by the camera device and pre-stored comparison pictures; identify human features in the pictures to be verified, wherein the human features include contour information, position information and human body part categories; create a cropping carrier based on the contour information and position information of the human features; use the cropping carrier to crop multiple pictures to be verified and comparison pictures to obtain multiple environment verification pictures and environment comparison pictures; compare the multiple environment verification pictures with the environment comparison pictures to calculate the similarity; when the similarity of the multiple environment verification pictures does not meet the threshold condition, output the detection result of forged face information; when the similarity of the multiple environment verification pictures meets the threshold condition, output the detection result of judging as liveness.

[0061] The principle of the present invention is that the camera device of the self-service correction terminal adopts a fixed position, and the background environment of the shooting area is unchanged. When the correction object operates in front of the self-service correction terminal, the body of the correction object will block the local area of ​​the background environment when it appears in the shooting area. By identifying the outline of the human body features, the picture to be verified and the comparison picture are cropped to obtain a picture with the human body part deleted. Under normal circumstances, the similarity of the two pictures with the human body part deleted is extremely high, and then by comparing the similarity between the two, it is judged whether the environment is abnormal. For example, when using an electronic screen to play a video for biometric attack, because the frame of the electronic screen and the background displayed by the electronic screen are different from the environment of the comparison picture, then through similarity comparison, it can be identified whether there is an attack on biometrics; or, when using a photo of a face print for biometrics, the background in the photo is different from the environment of the comparison picture, and then through similarity comparison, it can be identified whether there is an attack on biometrics.

[0062] The present invention only requires a single camera device, does not require other auxiliary equipment for liveness detection, and does not require the design of complex liveness detection algorithms to identify whether there is an attack on biometrics. It improves the security of the self-service correction terminal, prevents the correction object from making the liveness detection successful through attacks on biometrics, completes the comparison of biological information, and effectively and reliably reduces the phenomenon of correction objects being out of management. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] The specific embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings, wherein:

[0064] Figure 1 It is a schematic diagram of the functional device composition of the intelligent self-service correction terminal of the present invention;

[0065] Figure 2 It is a flow chart of the environment comparison process of the intelligent self-service correction terminal of the present invention;

[0066] Figure 3 It is a schematic diagram of the process of video decryption and uploading of the intelligent self-service correction terminal of the present invention;

[0067] Figure 4 It is a flow chart of the similarity calculation process of the intelligent self-service correction terminal of the present invention;

[0068] Figure 5 It is a flow chart of a liveness detection process based on a random behavior sequence of the intelligent self-service correction terminal of the present invention;

[0069] Figure 6 It is a schematic diagram of the positioning monitoring process of the intelligent self-service correction terminal of the present invention;

[0070] Figure 7 It is a complete flow chart of video encryption embedding hidden information of the present invention;

[0071] Figure 8 It is a schematic diagram of the image capture device of the intelligent self-service correction terminal of the present invention being extended and stored;

[0072] Fig. 9 is a side view of the intelligent self-service correction terminal of the present invention;

[0073] Fig.10 is a top view of the intelligent self-service correction terminal of the present invention;

[0074] Fig.11 is a bottom view of the intelligent self-service correction terminal of the present invention;

[0075] Fig.12 is a front view schematic diagram of a document placement pad of the present invention;

[0076] In the figure:

[0077] 10-terminal body, 11-base, 111-positioning part, 12-bracket body, 13-front display screen, 14-rear display screen;

[0078] 20- Image capture device;

[0079] 30- document placement pad, 31- positioning hole. DETAILED DESCRIPTION

[0080] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0081] Embodiment 1

[0082] The technical security loopholes of existing self-service correction terminals are mainly technical defects in liveness detection. The self-service correction terminal performs online identity authentication on the current person through a camera. Face liveness detection technology is a technology that determines whether the captured face is a real live face or a processed face image. Common attack methods against liveness detection include: photo attacks and video attacks (such as face images printed on color paper, digital face images displayed or played on the screen of an electronic device, and face images of people wearing makeup or masks, etc.).

[0083] However, with more and more attacks on biometrics, in order to solve security loopholes, avoid making liveness detection successful through abnormal channels, complete the comparison of biometric information, and complete the registration and learning without being present, etc., so that the correction object is out of management. The present invention provides an intelligent self-service correction terminal, which is equipped with an anti-counterfeiting detection device for face anti-fraud liveness detection, and the anti-counterfeiting detection device is configured to obtain multiple pictures to be verified collected by the camera device and pre-stored comparison pictures; identify human features in the pictures to be verified, and the human features include contour information, position information and human body part categories; create a cropping carrier based on the contour information and position information of the human features; use the cropping carrier to crop multiple pictures to be verified and comparison pictures to obtain multiple environment verification pictures and environment comparison pictures; compare multiple environment verification pictures with environment comparison pictures to calculate similarity; when the similarity of multiple environment verification pictures does not meet the threshold condition, output the detection result of forged face information; when the similarity of multiple environment verification pictures meets the threshold condition, output the detection result determined as liveness.

[0084] like Figure 1 As shown, an intelligent self-service correction terminal described in the present invention includes a terminal body, and the terminal body is provided with a camera device and an anti-counterfeiting detection device.

[0085] The camera device is used for liveness detection, identity verification, etc., and is also used to record and store the corresponding activities of the correction subject in the form of video data when the correction subject is performing operations such as reporting and learning. The camera device is fixedly installed on the body of the intelligent self-service correction terminal.

[0086] In one implementation, the camera device starts collecting video data of the target area when the correction subject requests authentication, or when the intelligent self-service correction terminal is awakened, or when the identity information collection device is awakened.

[0087] like Figure 2 As shown, it is a schematic diagram of the process flow of the anti-counterfeiting detection device executing the environment comparison process. Specifically, the anti-counterfeiting detection device is configured as follows:

[0088] S100: Acquire a plurality of pictures to be verified captured by the camera device and pre-stored comparison pictures.

[0089] In one implementation, when the correction subject uses a self-service correction terminal to perform management activities such as reporting for learning, based on the authentication request, reporting, learning sign-in, etc. initiated by the correction subject, the self-service correction terminal controls the camera device to collect multiple pictures to be verified.

[0090] In one implementation, the multiple images to be verified may be obtained by continuous shooting, and when the user initiates a request, the camera device continuously shoots multiple consecutive images to be verified. In another implementation, the multiple images to be verified may be obtained by interval shooting, and the self-service correction terminal controls the camera device to shoot multiple images to be verified that are discontinuous in time and have intervals according to a preset period.

[0091] It is preferred to use multiple pictures to be verified that are discontinuous in time and have intervals. If there is an abnormal statistical attack on biometric identification, it is often the case that the correction subject holds the corresponding electronic screen or photo in front of the self-service correction terminal for forged verification. By taking pictures at intervals, it is better to capture the attack vulnerabilities of the correction subject, such as the appearance of multiple different human features, or the proportion of human features in the background environment varies from time to time, so that when an abnormality is found, the relevant video can be checked.

[0092] In the present invention, the comparison picture is a picture of the background environment taken by the camera device of the self-service correction terminal. Since the camera device of the self-service correction terminal adopts a fixed position, the background environment of the area it shoots is unchanged. By using the picture of the background environment as the comparison picture, it is compared with the picture to be verified.

[0093] It should be noted that in order to improve the reliability and accuracy of the comparison, the comparison pictures are required to include only fixed objects, which are items whose positions and outline sizes remain unchanged in the background environment, such as green plants placed in the organization, billboards on wall panels, guide signs, etc.; generally, these fixed objects will not move.

[0094] In a preferred implementation, the comparison picture is obtained by automatically taking pictures when the self-service correction terminal is in standby mode. Standby mode means that the self-service correction terminal enters a dormant state when there is no user operation for a certain period of time. When the self-service correction terminal is in standby mode, the camera device is controlled to take multiple photos of the background environment at a preset period, and the object features in the multiple photos are compared to see if they are exactly the same. If so, they are used as comparison photos.

[0095] In another implementation, the comparison image may be taken by a person of the institution operating a self-service correction terminal and pre-stored in the self-service correction terminal.

[0096] S200: Identify human features in the image to be verified, where the human features include contour information, position information, and human body part categories.

[0097] In the present invention, human feature recognition and contour recognition of images are existing technologies in the art, such as human feature recognition technology based on deep learning, or building a deep neural network structure to predict the semantic category of an image at the pixel level. Those skilled in the art can implement it according to the description of this application, and no further explanation is given here.

[0098] In one implementation, the human body part categories mainly include head, neck, torso, limbs, and hands.

[0099] S300: Creating a cutting carrier based on the contour information and position information of the human body features.

[0100] The cropping carrier, as a carrier of the cropping template and the content to be cropped, can be used to set the picture as the cropping template and the picture as the content to be cropped.

[0101] S400: Using a cropping carrier to crop a plurality of to-be-verified pictures and comparison pictures to obtain a plurality of environment verification pictures and environment comparison pictures.

[0102] The cutting carrier is used as a cutting template, and the cutting area of ​​the cutting template is used as a cutting mold for the content to be cut. The content to be cut is cut according to the contour information of the cutting mold, and the part within the cutting mold range and the part outside the cutting mold range can be obtained. The part within the cutting mold range is the part that needs to be deleted, and the part outside the cutting mold range is the part that needs to be retained.

[0103] In one implementation, the image to be verified and the comparison image may be an image with an alpha channel, such as a PNG image, which may include a transparent portion and a non-transparent portion. After the image to be verified / the comparison image is cropped according to the cropping template, the cropping area is set to the transparent portion.

[0104] In one implementation, the image to be verified and the comparison image can be jpg images. After the image to be verified / the comparison image is cropped according to the cropping template, the cropping area is filled with a single pixel, such as white, to complete the cropping.

[0105] In a specific implementation, the method of using a cropping carrier to crop multiple images to be verified and comparison images to obtain multiple environment verification images and environment comparison images specifically includes:

[0106] S410: adding a cropping carrier to the image to be verified, and setting the cropping carrier to the content to be cropped on the image to be verified;

[0107] S420: Crop the image to be verified along the contour information in the cropping carrier according to the position information of the cropping carrier, and filter the portion of the image to be verified within the range surrounded by the contour information of the cropping carrier to obtain an environment verification image.

[0108] S430: adding a cropping carrier to the comparison image, and setting the cropping carrier to the content to be cropped on the comparison image;

[0109] S440: Crop the comparison image along the contour information of the cropping carrier according to the position information of the cropping carrier, filter the portion of the comparison image within the range surrounded by the contour information, and obtain the environment comparison image.

[0110] In one implementation, multiple images to be verified correspond to multiple cropping carriers. Since the correction object may move, the outline size and proportion of the correction object in the photo are not constant. To this end, the contour information and position information of the human features of multiple images to be verified are used to create cropping carriers in turn. Each cropping carrier has a corresponding image to be verified. Each cropping carrier is used to crop the image to be verified and the comparison image in turn to obtain multiple environmental verification images, and multiple corresponding environmental comparison images are obtained. One-to-one comparison is performed to improve the accuracy and reliability of the comparison.

[0111] S500: Compare the plurality of environment verification pictures with the environment comparison pictures, and calculate similarity.

[0112] S610: When the similarity of multiple environment verification pictures does not meet the threshold condition, output the detection result of the forged face information.

[0113] S620: When the similarity of multiple environment verification images meets the threshold condition, output the detection result of judging as a living body.

[0114] Based on steps S500, S610 and S620, the principle of the present invention is that the camera device of the self-service correction terminal adopts a fixed position, and the background environment of the shooting area is unchanged. When the correction object operates in front of the self-service correction terminal, the body of the correction object will block the local area of ​​the background environment when it appears in the shooting area. By identifying the outline of the human body features, the picture to be verified and the comparison picture are cropped to obtain a picture with the human body part deleted. Under normal circumstances, the similarity of the two pictures with the human body part deleted is extremely high, and then by comparing the similarity between the two, it is determined whether the environment is abnormal. For example, when using an electronic screen to play a video for biometric attack, because the frame of the electronic screen and the background displayed by the electronic screen are different from the environment of the comparison picture, then through similarity comparison, it can be identified whether there is an attack on biometrics; or, when using a photo printed with a face for biometrics, the background in the photo is different from the environment of the comparison picture, and then through similarity comparison, it can be identified whether there is an attack on biometrics.

[0115] The present invention only requires a single camera device, does not require other auxiliary equipment for liveness detection, and does not require the design of complex liveness detection algorithms to identify whether there is an attack on biometrics. It improves the security of the self-service correction terminal, prevents the correction object from making the liveness detection successful through attacks on biometrics, completes the comparison of biological information, and effectively and reliably reduces the phenomenon of correction objects being out of management.

[0116] Among them, the present invention provides a specific technical means for calculating similarity, using the Average Hash (aHash) algorithm. The design of the Average Hash algorithm is relatively simple. It mainly uses the low-frequency components of the image, removes the high-frequency components of the image by shrinking the image, retains the low-frequency information, and uses the image grayscale method to remove the image color to further remove the high-frequency components. On this basis, the pixel average of the grayscale image is calculated. Traverse each pixel of the grayscale image and compare it with the pixel mean. If it is greater than the mean, record 1; otherwise, record 0. The binary string obtained is the image hash value, also known as the image fingerprint.

[0117] like Figure 3 As shown, the comparison of the plurality of environment verification pictures with the environment comparison pictures and calculation of similarity specifically includes:

[0118] S510: reducing the size of the multiple images to be verified and the comparison images to 64*64.

[0119] S520: Convert the 64*64 to-be-verified image and the 64*64 comparison image into grayscale images, and obtain a plurality of to-be-verified grayscale images and comparison grayscale images.

[0120] S530: Calculate pixel averages of multiple grayscale images to be verified and comparison grayscale images.

[0121] S540: traverse each pixel of the plurality of grayscale images to be verified and the comparison grayscale images, compare the pixel with the pixel average value, and obtain hash values ​​of the plurality of images to be verified and the comparison images;

[0122] Specifically, if the pixel ≥ the pixel average, record 1, otherwise record 0, and the resulting binary string of bits is the hash value of the image.

[0123] S550: Calculate the Hamming distances between the multiple to-be-verified images and the comparison images in sequence to obtain similarities between the multiple to-be-verified images and the comparison images.

[0124] The smaller the Hamming distance between the multiple to-be-verified images and the comparison images, the more similar the images are, and the larger the distance, the greater the difference between the images. Based on the teachings of this application, those skilled in the art can set threshold conditions according to actual needs to judge the similarity of multiple environment verification images and output the detection results.

[0125] like Figure 4 As shown, in a preferred implementation, the anti-counterfeiting detection device is configured to perform the following steps before performing the environmental comparison process comparison:

[0126] S10: Identify whether a plurality of images to be verified contain an electronic screen;

[0127] S11: If not, then the environment comparison process is performed, and the anti-counterfeiting detection device executes the environment comparison process of steps S100-S620.

[0128] S12-1: If yes, identify the object features of the electronic screen, where the object features include the outline of the electronic screen and the position of the electronic screen in the image to be verified.

[0129] In one implementation, the type of the feature of the object can be identified by matching with preset identification features of various object features. For example, for an electronic screen, the preset identification features are various types of electronic product frames (such as mobile phones, tablet computers, laptops, etc.), moiré patterns, mirror reflection spots, electronic screen edges, etc.

[0130] In the present invention, the feature recognition and contour recognition of objects in images are existing technologies in the art, such as feature recognition technology based on deep learning, or building a deep neural network structure to predict the semantic category of an image at the pixel level. Those skilled in the art can implement it according to the description of this application, and no further explanation is given here.

[0131] S12-2: Determine whether the contour information of the human body feature is located within the electronic screen contour of the object feature.

[0132] S12-21: If yes, output the detection result of forged facial information; S12-22: If no, perform the environment comparison process.

[0133] Because the correctional subject may be holding an electronic device to watch the video when registering or studying, in order to reduce misjudgment, it is further determined whether the contour information of the human body feature is within the electronic screen contour of the object feature to determine whether there is a biometric attack on the video playback.

[0134] Embodiment 2

[0135] In the existing liveness detection process, when the correction object to be authenticated initiates the authentication request at the self-service correction terminal, the self-service correction terminal will issue the corresponding action command to obtain the face image or video, and perform liveness detection based on whether the acquired face image or video is consistent with the action behavior in the action command. For example, the user end prompts the action command through voice: blink, and when the user completes the blink, the liveness detection is passed.

[0136] The applicant found that the actions in the existing liveness detection method are relatively simple, and the order of the actions is fixed. The liveness detection can also be successful through pre-recorded videos, images, etc., to achieve an attack on the authentication request, so that the correction object can complete the registration learning through abnormal channels, etc. Therefore, the security of the existing liveness detection method is relatively low.

[0137] like Figure 1 As shown, the intelligent self-service correction terminal of the present invention also includes a liveness detection device, which is used to perform liveness detection on the correction object. Different facial movements can be randomly combined to further increase the difficulty of the action behavior, thereby improving the security of liveness authentication.

[0138] like Figure 5 As shown, the liveness detection device is configured to perform the following liveness detection process, including:

[0139] S1000: Receive an authentication request for an object to be authenticated and corrected.

[0140] In one implementation, when the correction subject to be authenticated logs in at the self-service correction terminal, it indicates that the correction subject has the intention to perform liveness detection. Therefore, when receiving the login request, it can be regarded as receiving the authentication request of the correction subject to be authenticated, thereby starting the execution of the liveness detection method. At the same time, the authentication request can be triggered during the process of checking in, learning, signing out, etc.

[0141] S2000: Generate at least two random behaviors; each of the random behaviors indicates a facial action, and different random behaviors indicate different facial actions.

[0142] In one implementation, different difficulty levels can be set according to the activities that the correction object needs to perform. For example, the number of random behaviors required to generate for login is 2. The number of random behaviors required to generate for the registration authentication request is 3; for another example, the number of random behaviors required to generate for the education and learning and sign-out authentication requests is 5.

[0143] Specifically, for different management activities, more random behaviors need to be generated. Since the more random behaviors there are, the more complex the random behavior sequence obtained by random combination is, the more difficult it is to be attacked during liveness detection, and the more secure liveness detection is.

[0144] Specifically, random behaviors may include blinking, raising the head, lowering the head, turning the head to the left, turning the head to the right, opening the mouth, closing the mouth, etc.

[0145] S3000: Randomly combine the at least two random behaviors to generate a random behavior sequence; and send the random behavior sequence to a display screen of the intelligent self-service correction terminal for display.

[0146] When the display screen of the intelligent self-service correction terminal displays a random behavior sequence, the camera device can be turned on to record the video. According to each facial action and its sequence in the received random behavior sequence, the correction subject performs each facial action in turn, thereby recording a video to be verified.

[0147] S4000: Acquire the video to be verified captured by the camera device; and obtain the facial action sequence of the correction object using the video to be verified.

[0148] Specifically, in order to ensure the continuity of the video and avoid counterfeit spliced ​​videos attacking the liveness detection method, after obtaining the video to be verified, the video to be verified can be tested for continuity. For example, the optical flow algorithm in the prior art can be used to determine the continuity of the video to be verified; or a hash algorithm can be used to calculate the similarity of the background image in each frame of the image to be identified after the video to be verified is framed, and the continuity of the video to be verified is determined when the similarity of the background image in each frame of the image to be identified reaches a preset threshold. Afterwards, if the continuity test passes, the subsequent steps can be continued.

[0149] S4100: Determine whether at least two facial actions included in the facial action sequence are the same as the facial actions indicated by at least two random actions in the random action sequence, and; determine whether the order of at least two facial actions included in the facial action sequence in the facial action sequence is the same as the order of at least two random actions in the random action sequence in the random action sequence;

[0150] S4200: If they are the same, then output the authentication success result. If not, then the liveness authentication fails.

[0151] Embodiment 3

[0152] At present, the positioning of correctional subjects is mainly managed by installing corresponding management programs on mobile phones, or wearing electronic products such as positioning bracelets. When correctional subjects enter judicial institutions to report for study, etc., positioning is used to determine whether correctional personnel are using self-service correctional terminals, and whether there are any acts of impersonation or escape from management. This method is convenient to use, but at the same time, there are corresponding security loopholes, such as positioning information can be forged, or self-service correction terminals can be moved to the side of correctional subjects through improper means to escape management, and then it is impossible to monitor whether correctional subjects report for study according to regulations.

[0153] like Figure 1 As shown, in a preferred implementation, the intelligent self-service correction terminal also includes a positioning device and a position monitoring device, and the positioning device is used to collect the current position information of the intelligent self-service correction terminal.

[0154] like Figure 6 As shown, the position monitoring device is configured to perform the following positioning monitoring process, including:

[0155] S5000: Obtain the current location information of the positioning device.

[0156] In one implementation, the positioning device may be a GPS module or other software or hardware with positioning functions.

[0157] S6000: Load the pre-stored geo-fence and determine whether the current location information is within the geo-fence.

[0158] Geo-fencing is a new application of LBS, which is to use a virtual fence to enclose a virtual geographic boundary. Geographic fences (i.e. a set of closed latitude and longitude coordinate pairs) are used to divide the area or street where the institution is located. When applied to community corrections, geo-fencing can be set for specific areas or streets for different institution addresses, so that it can be quickly and efficiently determined whether a certain point is within the fence.

[0159] In one implementation, the ray method can be used to determine whether a point is within the region. The ray method is a horizontal scan line method or a vertical line method. If there is a position point P (x, y), to determine whether it is within the polygon, a horizontal scan line (i.e., ray) can be drawn from the question point to the left. The number of intersections c between this line segment and the region boundary is calculated. If c is an odd number, the position point is within the polygon; if it is an even number, the position point is outside the polygon.

[0160] S6100: If yes, return the hit result, and instruct the intelligent self-service correction terminal to perform the liveness detection process based on the hit result;

[0161] S6200: If not, a miss result is returned, and an abnormal location alarm message is generated and issued based on the miss result, and the external service of the intelligent self-service correction terminal is stopped.

[0162] In a specific embodiment, the abnormal location alarm information is sent to the manager's smart terminal in the form of a text message.

[0163] Through this design, before the correctional object reports for study, the self-service correction terminal pre-determines the location of the terminal to eliminate the problem that the autonomous correction terminal is moved and helps the correctional object escape management, and then monitors whether the correctional object reports for study according to the regulations; through the coordination of anti-liveness detection attacks and self-service correction terminal positioning and identification, the security of information management of correctional objects is comprehensively improved.

[0164] Embodiment 4

[0165] The existing self-service correction terminals have certain security loopholes. Computer technology can be used to create multimedia data to replace the data stored in the self-service correction terminals. Figure 1 and Figure 7 As shown, the intelligent self-service correction terminal described in the present invention includes an identity information collection device, a camera device and a data processing device.

[0166] Among them, the identity information collection device is mainly used for identity authentication. The identity information collection device is used to collect the identity information of the correction object, such as for collecting ID card detection, face recognition and biometric fingerprints, and verifying the identity information to determine whether the currently collected identity information matches the correction object. Identity authentication is required when the correction object reports, studies, and other operations. The identity information collection device is implemented by the relevant hardware set on the intelligent self-service correction terminal and the software program deployed on the intelligent self-service correction terminal. It is common knowledge in the field and will not be explained in detail here.

[0167] The data processing device is connected to the camera device, and the data processing device is configured to encrypt the video data of the camera device, including the following encryption method:

[0168] S1: Acquire the video data currently collected by the camera device and the identity information of the correction object currently collected by the identity information collection device.

[0169] In one implementation, when the correction subject is performing operations such as reporting in or learning, the camera device records the correction subject's activities and forms video data. The video data is preferably encoded using H.264 or HEVC standards, or H.264 / HEVC standards.

[0170] S2: Encrypt the video data, and store the encrypted video data locally in association with the identity information.

[0171] Through this design, the video is encrypted and stored inside the self-service correction terminal to distinguish the video collected by the self-service correction terminal itself from the forged video stored in the self-service correction terminal through abnormal means, and then it can be identified whether the correction object is evading management, so as to solve the security loopholes of the existing technology and have high security.

[0172] In a specific implementation, the encrypting the video data specifically includes: after entropy decoding the video data bitstream, key encrypting the non-zero coefficients of the syntax element DCT coefficients, and then obtaining the encrypted video data after entropy encoding.

[0173] It should be noted that the entropy coding of the HEVC coding standard adopts CABAC technology. The CABAC entropy coding process is divided into two steps: binarization of syntax elements and arithmetic coding of binary bit streams (bins). Arithmetic coding is further divided into conventional arithmetic coding based on context model and bypass arithmetic coding using approximately equal probability model.

[0174] Specifically, the present invention implements video encryption by operating the sign bits of non-zero coefficients of syntax element DCT coefficients after entropy decoding the video bit stream encoded in H.264 or HEVC standard.

[0175] In the CABAC entropy coding of the present invention, the syntax element using the bypass coding mode includes the sign bit of the non-zero coefficient of the syntax element after DCT transformation. Changing its sign bit can produce obvious perceptual disturbance effect on the reconstruction of the video collected by the self-service correction terminal, but has no effect on the bitstream structure.

[0176] In a specific implementation manner, the step of encrypting the video data in S200 specifically includes:

[0177] S2-1: Generate a binary pseudo-random sequence using an encryption key. S2-2: Obtain a binary string of the original video bitstream captured by the camera device through CABAC entropy decoding. S2-3: Perform bitwise XOR operation on the binary string of the non-zero coefficient sign bits of the syntax element DCT coefficients to obtain an encrypted binary string. S2-4: Perform entropy encoding on the encrypted binary string. S2-5: Repeat the above 230 to S240 until the video bitstream ends and the encrypted video data is obtained.

[0178] In a preferred implementation, the data processing device is further configured to decrypt the encrypted video data, including the following decryption method:

[0179] S3: When a data upload event is triggered, identify whether the video data stored locally and to be uploaded is encrypted.

[0180] In one example, the triggering of the data upload event can be an active triggering by a manager, or it can be automatically triggered when the self-service correction terminal is idle, or it can be automatically triggered according to a preset period, or it can be automatically triggered when the number threshold or storage size threshold of the locally stored video meets. The identification of encrypted videos is a prior art in this field and will not be described in detail here.

[0181] S3-1: If yes, decrypt the video data stored locally, and upload the decrypted video data and identity information based on the communication link.

[0182] In a specific implementation, the decrypting of the locally stored video data specifically includes: entropy decoding the encrypted video data code stream and then decrypting it using a key, and then entropy encoding it to obtain the decrypted video data.

[0183] In a specific implementation, the data processing device has the authorization of the encryption key, uses the encryption key to first decrypt the uploaded video data, utilizes the reversibility of the XOR operation to decrypt the uploaded video data stream, and restores the original binary string to obtain the decrypted video data.

[0184] S3-2: If not, generate and issue a video forgery alarm message.

[0185] In one implementation, the video forgery alarm information is sent to the instructor's mobile terminal via text message, or the self-service correction terminal emits an audible and visual alarm light, or the corresponding alarm information is displayed on the display screen of the self-service correction terminal.

[0186] Through this design, when a data upload event is triggered, the self-service correction terminal identifies whether the data video stored locally and to be uploaded is encrypted. Only the encrypted video is decrypted and transmitted, thereby avoiding the transmission of forged videos to the background and further improving security.

[0187] The self-service correction terminal of the present invention encrypts and stores the video, and can transmit it to the background after decrypting the video. The two work together to improve the security of information management of the correction object through encryption and decryption technology.

[0188] In a preferred implementation, based on the third embodiment, the data processing device is further configured as follows:

[0189] S5: Obtain the current position information of the positioning device. S6: Perform AES-OFB encryption on the current position information by using the embedded key to obtain a binary sequence to be embedded. S7: Perform entropy decoding on the encrypted video data bitstream to obtain a binary string of encrypted video data. S8: Replace the target position in the binary string with the binary sequence, perform entropy encoding on the replaced binary string, and obtain encrypted video data containing the current position information.

[0190] In a specific implementation, since information hiding is applied to an encrypted video, data embedding should not affect the encrypted sign bit. In a specific implementation, the binary string of MVD can also be selected as the information hiding domain. The current position information can be hidden and stored in the video data by applying the binary string codeword equal length replacement method to the MVD suffix part. Encryption and information hiding are both completed in the CABAC bypass coding mode, and the "current position information" mark can be directly embedded in the encrypted video.

[0191] This design further increases the difficulty of video forgery and prevents video data from being forged, thereby improving security. On the other hand, by embedding the current location information into the video data, when there are disputes or problems with the video data, the location information of the video data when it was shot can be obtained through reverse data extraction. Based on the location information, it can be determined whether the video was shot within the corresponding geographic fence, and then whether there is a problem of forged video. This can also provide evidence for the authenticity and reliability of the video data.

[0192] Embodiment 5

[0193] like Figure 8 to Figure 12 As shown, the preferred structure of the intelligent self-service correction terminal described in the present invention.

[0194] like Figure 8 As shown, the intelligent self-service correction terminal of the present invention includes a terminal body 10, and an image capture device 20 movably connected to the terminal body 10. The image capture device 20 is used for image acquisition of document scanning; the image capture device 20 can be extended from the terminal body 10 and can be suspended outside the side wall of the terminal body 10, and the image capture device 20 can also be stored in the terminal body.

[0195] Through this design, the self-service correction terminal of the present application utilizes the external space of the terminal body as the document scanning area, and thus does not need to add an additional scanning space for placing documents at the bottom of the self-service correction terminal, so as to reduce the volume of the self-service correction terminal, facilitate the storage and transportation of the self-service correction terminal, and reduce the material and weight. On the other hand, the image capture device can be stored when not in use, and has storage performance while ensuring the document scanning function, and has good protection for the core components of the image capture device.

[0196] In a specific implementation, Fig. 9 As shown, the terminal body includes a base 11, a bracket body 12 installed vertically on the base 11, a front display screen 13 and a rear display screen 14 for human-computer interaction, and electrical components for realizing multiple functions.

[0197] In one implementation, the electrical components include a controller composed of a single-chip microcomputer or a control circuit, a fingerprint scanning device, an ID card recognition device and the above-mentioned image shooting device, a camera device for shooting faces for face recognition, a speaker, etc. The electrical components endow the self-service correction terminal with integrated voice recognition, face recognition, video recording, ID card recognition, fingerprint recognition and document scanning functions. The speaker can also realize publicity broadcasts, legal services and convenient services. The electrical components can also communicate and transmit data with external devices through a dedicated network, reduce the burden on community correction staff, and improve administrative work efficiency.

[0198] In a specific implementation, the base 11 and the bracket body 12 are both made of plastic shells, the front display screen 13 and the rear display screen 14 are installed on the front and back of the bracket body 12 respectively, and the image capture device 20 is movably connected to the bracket body 12, and the image capture device 20 is located between the front display screen 13 and the rear display screen 14.

[0199] Specifically, the front display screen and the rear display screen can be pivotally connected or fixedly connected to the top of the bracket body. The camera device is fixedly connected to the bracket body, the ID card recognition device and the fingerprint scanning device are integrated on the base, and a mark for placing an ID card area or a finger area is set on the upper surface of the base.

[0200] In a preferred implementation, when the image capture device 20 swings to the target position for document scanning, the image capture device 20 is suspended outside the upper side wall of the terminal body 10. With this configuration, the image capture device 20 is suspended outside the side wall of the upper part of the terminal body 10 (preferably the top position of the bracket body), so that the image capture device and the desktop are separated by a suitable height in the vertical direction, which meets the focus requirements of the image capture device, and enables the image capture device to capture the entire document to avoid missing parts, making it more convenient for personnel to operate, and reducing the operation of aligning the image capture device with the shooting area when placing documents.

[0201] Preferably, when the image capture device 20 swings to the target position for document scanning, the image capture device 20 is horizontally suspended outside the upper side wall of the terminal body 10, so that the optical lens of the image capture device 20 is vertically facing the document located below. Regarding the movable connection between the image capture device 20 and the terminal body 10, the image capture device 20 can be extended from the terminal body 10 and suspended outside the side wall of the terminal body 10, and the image capture device 20 can also be stored in the terminal body.

[0202] Specifically, the image capture device 20 is pivotally connected to the terminal body 10 to form a rotating pair; the image capture device 20 can swing outward from the side wall of the terminal body 10 around the pivot, and can also be stored and attached to the side wall of the terminal body 10.

[0203] In one implementation, Figure 8As shown, one side of the image capture device 20 is pivotally connected to the bracket body 12 of the terminal body 10, and the image capture device 20 is horizontally stored on the upper surface of the bracket body 12, between the front display screen 13 and the rear display screen 14. When in use, the image capture device 20 is flipped outward from the upper surface of the terminal body 10 until the image capture device 20 is horizontally suspended outside the side wall of the terminal body 10; when stored, the image capture device 20 is flipped back to the upper surface of the bracket body 12 and placed.

[0204] In another implementation, Figure 8 As shown, one side of the image capture device 20 is pivotally connected to the terminal body 10. When the image capture device 20 is stored, it is vertically attached to the side wall of the bracket body 12 and is located between the front display screen 13 and the rear display screen 14. When in use, the image capture device 20 is flipped outward from the side wall of the terminal body 10 from bottom to top until the image capture device 20 is horizontally suspended outside the side wall of the terminal body 10; when stored, the image capture device 20 is flipped back to the side wall of the bracket body 12 and placed.

[0205] In a preferred embodiment, a limiter assembly may be provided on the pivot structure between the image capture device 20 and the terminal body 10 to prevent the image capture device from swinging excessively and from falling downward due to gravity. The limiter assembly may be a damping assembly provided on the pivot structure, so that the image capture device can be suspended at will; or it may be a raised support member provided on the side wall of the bracket body, such as a rotating locking structure of a flip-up desk lamp. The limiter assembly can be implemented by a person skilled in the art according to the description of this application, and is not limited to the implementation method and specific structure described in this application.

[0206] In a preferred embodiment, when the image capture device swings to a target position for document scanning, the image capture device is horizontally suspended outside the side wall of the terminal body. The limit assembly can also be used to limit the swing of the image capture device and make the image capture device hover at the target position.

[0207] In a preferred implementation, Fig.12 As shown, in order to facilitate the user to quickly and accurately place the document to complete the document scanning work, the location where the document is placed corresponds to the shooting area of ​​the image shooting device. The self-service correction terminal also includes a document placement pad 30 detachably connected to the terminal body.

[0208] In a preferred embodiment, a document placement area mark 32 is provided on the upper surface of the document placement pad 30 ; when the image capture device 20 swings to a target position for document scanning, the image capture device 20 correspondingly hovers above the document placement pad 30 .

[0209] Specifically, the document placement area mark 32 may include an ID card placement area mark, an A6 paper grid paper area mark, an A5 paper placement area mark, and an A4 paper placement area mark, etc. Those skilled in the art may set it according to actual needs. The document placement area mark is a pattern set on the document placement pad, and the color of the document placement area mark has a greater contrast with the main body color of the document placement pad.

[0210] Specifically, at least two positioning parts 111 are disposed at the lower part of the terminal body; the document placement pad 30 is provided with positioning holes 31 adapted to the plurality of positioning parts, and the document placement pad 30 is sleeved with the positioning parts through the positioning holes 31 to be connected to the terminal body.

[0211] When in use, you only need to lift the desktop self-service correction terminal, align the document placement pad on the desktop with the positioning part on the base, put down the self-service correction terminal and press the document placement pad. It is easy to use, has a long service life, is not easy to damage, has a simple structure, and reduces the cost of use.

[0212] Preferably, the document placement pad 30 is a rectangular pad made of a flexible plastic material, such as silicone or PVC.

[0213] In a specific implementation, the image capture device 20 includes a long block-shaped shell and a camera assembly disposed in the shell; the optical lens of the camera assembly is exposed on the outer surface of the shell.

[0214] By using a high-speed scanner, it can complete a high-speed scan in one second. It has an OCR text recognition function and can convert scanned images into editable word documents. It can also take photos, record videos, copy, fax online paperless, make e-books, trim and straighten edges. It is a perfect solution to make office work easier, faster and more environmentally friendly. The high-speed scanner uses a USB2.0 interface with a transmission speed of up to 480Mbps. It uses a USB3.0 interface with a transmission speed of up to 5Gb / s. The 200\300\500\1000\1400\1600 megapixel sensor is equipped with a 300\500\800\1000\1400\1600 megapixel high-definition lens, providing high-quality scanning, with a maximum scanning size of up to A3 format. Whether it is a color book, a bill ID card or a manuscript file, you can easily get JPG or save it in a set format.

[0215] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Therefore, any modification, equivalent change and modification made to the above embodiment according to the technical essence of the present invention without departing from the technical solution of the present invention shall still fall within the scope of the technical solution of the present invention.

Claims

1. An intelligent self-service correction terminal, characterized in that: The terminal comprises a terminal body, the terminal body is provided with a camera device for liveness detection and recording the activities of the correction object; the terminal body is also provided with an anti-counterfeiting detection device connected to the camera device, and the anti-counterfeiting detection device is configured as follows: Acquire multiple pictures to be verified captured by the camera device and pre-stored comparison pictures; Identify human features in the image to be verified, where the human features include contour information, position information, and human body part categories; Creating a cutting vector based on the contour information and position information of human body features; Using a cropping carrier to crop multiple to-be-verified pictures and comparison pictures, a plurality of environment verification pictures and environment comparison pictures are obtained; Compare the plurality of environment verification pictures with the environment comparison pictures, and calculate the similarity; When the similarity of multiple environment verification images does not meet the threshold condition, the detection result of forged face information is output; When the similarity of multiple environmental verification images meets the threshold condition, the detection result of judging as living body is output; The anti-counterfeiting detection device is configured to perform the following steps before performing the environmental comparison process comparison: Identify whether multiple images to be verified contain electronic screens; If not, then the environmental comparison process is carried out; If yes, identify the object features of the electronic screen, the object features including the outline of the electronic screen and the position of the electronic screen in the image to be verified; Determining whether the contour information of the human body feature is located within the electronic screen contour of the object feature; If yes, output the detection result of forged face information; If not, the environment comparison process is performed.

2. The intelligent self-service correction terminal according to claim 1, characterized in that: The method of using a cutting carrier to cut a plurality of images to be verified and comparison images to obtain a plurality of environment verification images and environment comparison images specifically includes: Adding a cropping carrier to the image to be verified, and setting the cropping carrier to be the content to be cropped on the image to be verified; According to the position information of the cropping carrier, the image to be verified is cropped along the contour information in the cropping carrier, and a portion of the image to be verified within a range surrounded by the contour information of the cropping carrier is filtered to obtain an environment verification image; Adding a cropping carrier to the comparison image, and setting the cropping carrier to be the content to be cropped on the comparison image; The comparison picture is cropped along the contour information of the cropping carrier according to the position information of the cropping carrier, and the part of the comparison picture within the range surrounded by the contour information is filtered to obtain the environment comparison picture.

3. The intelligent self-service correction terminal according to claim 2, characterized in that: The step of comparing the plurality of environment verification pictures with the environment comparison pictures and calculating the similarity specifically includes: Reduce the size of multiple images to be verified and compared to 64*64; Convert the reduced image to be verified and the comparison image into grayscale images to obtain multiple grayscale images to be verified and comparison grayscale images; Calculate the pixel averages of multiple grayscale images to be verified and the comparison grayscale images; Traverse each pixel of multiple grayscale images to be verified and compared grayscale images, compare the pixel with the pixel average value, and obtain hash values ​​of multiple images to be verified and compared images; The Hamming distances between the multiple to-be-verified images and the comparison images are calculated in sequence to obtain the similarities between the multiple to-be-verified images and the comparison images.

4. The intelligent self-service correction terminal according to claim 1, characterized in that: The intelligent self-service correction terminal also includes a liveness detection device, which is configured as follows: Receiving a certification request for a subject to be certified and corrected; generating at least two random behaviors; each of the random behaviors indicates a facial action, and different random behaviors indicate different facial actions; Randomly combining the at least two random behaviors to generate a random behavior sequence; sending the random behavior sequence to a display screen of the intelligent self-service correction terminal for display; Acquire the video to be verified captured by the camera device; Using the video to be verified, the facial action sequence of the correction object is obtained; determining whether at least two facial actions included in the facial action sequence are the same as facial actions indicated by at least two random behaviors in the random behavior sequence, and; Determining whether the order of at least two facial actions included in the facial action sequence in the facial action sequence is the same as the order of at least two random actions in the random action sequence in the random action sequence; If they are the same, the authentication success result is output.

5. The intelligent self-service correction terminal according to claim 1, characterized in that: The intelligent self-service correction terminal also includes a positioning device and a position monitoring device. The positioning device is used to collect the current position information of the intelligent self-service correction terminal. The position monitoring device is configured as follows: Obtaining current location information of the positioning device; Load the pre-stored geo-fence and determine whether the current location information is within the geo-fence; If yes, a hit result is returned, and based on the hit result, the intelligent self-service correction terminal is instructed to perform a liveness detection process; If not, a miss result is returned, and an abnormal position alarm message is generated and issued based on the miss result, and the external service of the intelligent self-service correction terminal is stopped.

6. The intelligent self-service correction terminal according to claim 1, characterized in that: The intelligent self-service correction terminal also includes an identity information collection device and a data processing device, the data processing device is connected to the camera device, and the data processing device is configured to encrypt the video data of the camera device, including the following encryption method: Acquire the video data currently collected by the camera device and the identity information of the correction object currently collected by the identity information collection device; Encrypting the video data, and storing the encrypted video data locally in association with the identity information; The data processing device is further configured to decrypt the encrypted video data, including the following decryption method: When a data upload event is triggered, identify whether the video data stored locally and to be uploaded is encrypted; If yes, decrypt the video data stored locally, and upload the decrypted video data and identity information based on the communication link; If not, a video forgery alarm message is generated and issued.

7. The intelligent self-service correction terminal according to claim 1, characterized in that: The terminal body is provided with an image capturing device for document scanning; the image capturing device is pivotally connected to the terminal body to form a rotating pair; The image capture device can swing outward from the side wall of the terminal body around the pivot point, and can also be stored and attached to the side wall of the terminal body.

8. The intelligent self-service correction terminal according to claim 7, characterized in that: Also included is a document placement pad detachably connected to the terminal body; The upper surface of the document placement pad is provided with a document placement area mark; when the image shooting device swings to a target position for document scanning, the image shooting device correspondingly hovers above the document placement pad.

9. The intelligent self-service correction terminal according to claim 8, characterized in that: The terminal body comprises: A base, on which a bracket body installed upright is provided; A front display screen and a rear display screen for human-computer interaction, wherein the front display screen and the rear display screen are installed on the front and back of the bracket body respectively; A camera device for photographing a human face, wherein the camera device is fixedly connected to the support body; The image capture device is movably connected to the bracket body, and the image capture device is located between the front display screen and the rear display screen; an ID card recognition device and a fingerprint scanning device are integrated and installed on the base.

Images