Medical face recognition and treatment matching system, method and device

Abstract

The invention discloses a medical face recognition and treatment matching system, method and device, and the method comprises the steps: obtaining a reference face image of each patient, obtaining the reference face feature information of each patient, and storing the reference face feature information and original identity information of each patient; the method comprises the following steps: intercepting multiple frames of images from a video stream of a patient currently treated, acquiring real-time face images of the patient currently treated, acquiring facial features in each image in the real-time face images, and integrating the facial features of all the images to obtain instant feature data of the patient currently treated; on the basis of a comparison result of the instant feature data and the reference facial feature information, acquiring original identity information of the current treated patient; obtaining two-dimensional code information on the treatment sheet, and identifying second identity information contained in a two-dimensional code; comparing the original identity information with the second identity information, and performing treatment after the information is completely matched; the identity of the patient can be conveniently recognized, and the treatment process of the patient is matched and managed.

Description

Technical Field

[0001] The present invention relates to the field of face recognition technology, and in particular to a medical face recognition and treatment matching system, method and device. Background Art

[0002] Nowadays, when a patient is hospitalized, after entering the patient's identity information during the check-in stage, a wristband will be issued to the patient. The wristband contains the patient's name, medical record number, QR code, etc. When receiving treatment, nurses mostly use PDAs to scan the QR code on the wristband to identify the patient's identity, and scan the QR code on the treatment form to determine the corresponding patient identity. Only when the two identities are the same will the patient be given treatment.

[0003] However, for patients in the psychiatric department, their poor compliance in wearing wristbands and poor cooperation during treatment make it difficult to confirm and identify the patients. Summary of the invention

[0004] The purpose of the present invention is to provide a medical face recognition and treatment matching system, method and device to solve the technical problem in the prior art that patients have poor compliance in wearing wristbands and poor cooperation in performing treatment, which makes it difficult to confirm and identify the patient's identity.

[0005] In order to solve the above technical problems, the present invention specifically provides the following technical solutions:

[0006] A medical face recognition and treatment matching system, comprising a camera acquisition component, a data processing system, and a central processing system in communication with the data processing system;

[0007] The video acquisition component is used to obtain the instant facial image video stream of each patient and the QR code information on the treatment order;

[0008] The data processing system includes a data storage module, an image processing module and a comparison and recognition module;

[0009] The data storage module stores a reference face image of each patient, reference facial feature information of each patient acquired by the image processing module, and identity information of each patient;

[0010] The image processing module sequentially performs image frame splitting, image frame screening and image frame analysis on the real-time face image video stream actually acquired by the camera acquisition component, and acquires the real-time feature data of the real-time face image of the current patient in real time, and the image processing module is used to perform image decoding and analysis on the two-dimensional code information on the treatment order to identify the identity information contained in the two-dimensional code;

[0011] The comparison and recognition module compares the instant feature data of the real-time facial image of the current patient with the reference facial feature information of all patients to obtain the identity information corresponding to the current patient, and compares the identity information corresponding to the current patient with the identity information contained in the QR code to match the drug with the patient;

[0012] The data processing system sends the reference facial image, identity information of each patient and the matching information of the medicine and the patient during each treatment to the central processing system for data storage.

[0013] As a preferred solution of the present invention, when the image processing module sequentially performs image frame splitting and image frame screening on the instant facial image video stream actually acquired by the camera acquisition component, the instant facial image video stream is first split into multiple frames of facial images, and a clear image of each frame of facial image is used as the first screening condition, and the expression state of the screened facial image is used as the second screening condition to select a real-time facial image for identifying the instant feature data of the current patient.

[0014] In order to solve the above technical problems, the present invention also provides a medical face recognition and treatment matching method, comprising the following steps:

[0015] Acquire a reference facial image of each patient, process the reference facial image to obtain reference facial feature information of the patient, and save the reference facial feature information and original identity information of each patient;

[0016] A plurality of frames of images are intercepted from a video stream of a patient currently undergoing treatment to obtain a real-time facial image of the patient currently undergoing treatment, and a facial feature recognition algorithm is used to obtain facial features in each image of the real-time facial image, and the facial features of all images are integrated to obtain instant feature data of the patient currently undergoing treatment, and the original identity information of the patient currently undergoing treatment is obtained based on a comparison result between the instant feature data and the reference facial feature information;

[0017] Obtain the QR code information on the treatment order and identify the second identity information contained in the QR code;

[0018] Compare the original identity information with the second identity information, and only proceed with treatment when the information matches completely.

[0019] As a preferred solution of the present invention, the specific implementation method of obtaining the patient's baseline facial feature information is:

[0020] Performing image processing on the reference face image to extract facial features and facial curves of the patient, taking the center position of the patient's nostril as a central base point, and constructing a two-dimensional coordinate system according to the central base point;

[0021] Feature base points are selected from facial features and facial curves to form a facial base point polyline graph, and reference facial feature information is calculated respectively, wherein the reference facial feature information is divided into a left face facial feature group and a right face facial feature group.

[0022] As a preferred solution of the present invention, the characteristic elements included in the baseline facial feature information are: the distance between the central base point and the nose-eye line between the front ends of the two eyes, the angle between the two nose-eye lines, the distance between the two sides of the nose wing and the central base point, the distance between the corners of the mouth on both sides, the angle between the corners of the mouth on both sides and the central base point, the upper and lower spacing of the curve where the mouth is located, and the distance between the central base point and each inflection point of the facial curve.

[0023] As a preferred solution of the present invention, multiple frames of images are intercepted from the video stream, and the multiple frames of images are first screened using clarity as a screening condition to screen out sequence face images whose clarity is greater than a set threshold;

[0024] Performing secondary screening on the sequence of facial images with facial status as a screening condition to screen out real-time facial images whose facial status meets the set expectations, taking the real-time facial images as processing objects, and obtaining instant feature data of the patient currently undergoing treatment from the real-time facial images;

[0025] The instant feature data of the real-time facial image are integrated to form a set consisting of multiple feature elements.

[0026] As a preferred solution of the present invention, the method for performing secondary screening on the sequence of face images with facial status as a screening condition to screen out real-time face images whose facial status meets the set expectations is as follows:

[0027] Obtaining a mouth curve formed by the mouth of each of the sequence of facial images;

[0028] Taking the center base point of the two-dimensional coordinate system as the origin, selecting multiple comparison base points from the mouth curve, and determining the two-dimensional coordinate values ​​corresponding to the comparison base points, and determining the comparison base points representing the corners of the mouth on both sides based on the absolute values ​​of the X coordinates in all the two-dimensional coordinate values;

[0029] Determine a comparison base point representing the lowest position of the mouth based on the absolute value of the Y coordinate in all the two-dimensional coordinate values;

[0030] Calculate the facial state angle between the selected comparison base points representing the corners of the mouth on both sides and the comparison base point representing the lowest position of the mouth;

[0031] Select the sequence of face images whose facial state angles meet the set expectations as real-time face images.

[0032] As a preferred solution of the present invention, when the real-time facial image is processed to obtain the instant feature data of the patient currently undergoing treatment, the instant feature data is divided into a left face instant feature data group and a right face instant feature data group;

[0033] Compare the same feature elements in the left face instant feature data group with the left face facial feature groups of all reference facial images, and compare the same feature elements in the right face instant feature data group with the right face facial feature groups of all reference facial images, and determine the identity information of the current patient by combining the left face comparison matching rate and the right face comparison matching rate;

[0034] When the left face comparison matching rate is greater than the set threshold, and the right face comparison matching rate is greater than the set threshold, it is considered that the real-time face image is the same as the reference face image, and the identity information of the current patient is determined;

[0035] When the left face comparison matching rate is less than a set threshold, or the right face comparison matching rate is less than a set threshold, it is considered that the real-time face image is different from the reference face image.

[0036] As a preferred solution of the present invention, after obtaining the real-time feature data of the patient currently undergoing treatment, the number of feature elements in the left face real-time feature data group is compared with the number of feature elements in the right face real-time feature data group. When the number of feature elements in the left face real-time feature data group is not equal to the number of feature elements in the right face real-time feature data group, unilateral facial feature comparison is performed, and the set threshold of the left face comparison matching rate or the right face comparison matching rate is adjusted to increase.

[0037] As a preferred solution of the present invention, the original identity information of the patient currently undergoing treatment is extracted as the medical record number, the second identity information contained in the QR code is identified as the medical record number, the medical record numbers of the original identity information and the second identity information are compared in sequence, and a voice prompt is issued when the medical record numbers of the original identity information and the second identity information are exactly the same.

[0038] Compared with the prior art, the present invention has the following beneficial effects:

[0039] Before the specific treatment is carried out, the present invention uses a camera and an image processing module and adopts a face recognition method to automatically obtain the identity information of the current patient, and then uses the camera and the image processing module to identify the identity information contained in the QR code. The treatment is carried out only when the information is completely matched. This can ensure the correspondence between the medicine and the patient when the nurse performs the treatment, and prevent errors in the treatment.

[0040] In addition, when the present embodiment utilizes facial recognition to determine the patient's identity information, it specifically collects the facial video stream of the current patient, processes all frame images of the facial video stream in sequence to obtain a real-time facial image with high clarity and consistent with the expression state of the reference facial image, obtains the instant feature data of each image in the real-time facial image in sequence, and finally integrates all the instant feature data after error correction to obtain the instant feature data of the patient currently undergoing treatment. That is, before facial feature matching, the accuracy of the instant feature data of the patient currently undergoing treatment is improved by pre-processing the acquired facial image of the current patient, thereby reducing the feature requirements for facial feature matching. Using a fuzzy algorithm for face matching can shorten the recognition time and improve the recognition accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the implementation methods of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for the implementation methods or the description of the prior art. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other implementation drawings can be derived from the provided drawings without creative work.

[0042] Figure 1 is a structural block diagram of a face recognition and treatment matching system according to an embodiment of the present invention;

[0043] Figure 2 A schematic diagram of a flow chart of a face recognition and treatment matching method according to an embodiment of the present invention;

[0044] Figure 3 This is a schematic diagram of the appearance of a face recognition and treatment matching device according to an embodiment of the present invention;

[0045] Figure 4 A system diagram of a face recognition and treatment matching device according to an embodiment of the present invention

[0046] The numbers in the figure represent the following:

[0047] 1- Video acquisition component; 2- USB interface; 3- PDA shell; 4- PDA integrated system;

[0048] 41 - touch screen display module; 42 - communication module; 43 - battery module; 44 - voice broadcast module; 45 - data processing system; 46 - data storage module. DETAILED DESCRIPTION

[0049] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0050] like Figure 1 As shown, the present invention provides a medical face recognition and treatment matching system, including a camera acquisition component, a data processing system, and a central processing system communicatively connected to the data processing system.

[0051] The video acquisition component is used to obtain the real-time facial image video stream of each patient and the QR code information on the treatment order.

[0052] The data processing system includes a data storage module, an image processing module and a comparison and recognition module.

[0053] The data storage module stores the reference face image of each patient, the reference facial feature information of each patient obtained by the image processing module, and the identity information of each patient.

[0054] The image processing module performs image frame splitting, image frame filtering and image frame analysis on the real-time facial image video stream actually obtained by the camera acquisition component, and obtains the real-time feature data of the real-time facial image in real time. The image processing module is used to decode and analyze the QR code information on the treatment form to identify the identity information contained in the QR code.

[0055] The comparison and recognition module compares the instant feature data of the real-time facial image of the current patient with the baseline facial feature information of all patients to obtain the identity information corresponding to the current patient, and compares the identity information corresponding to the current patient with the identity information contained in the QR code to match the drug and the patient.

[0056] The data processing system sends each patient's reference facial image, identity information, and matching information between medicine and patient during each treatment to the central processing system for data storage.

[0057] In this embodiment, the medical face recognition and treatment matching system is mainly used in the inpatient treatment stage. When processing the electronic medical records of each patient, the camera acquisition component can be used to capture the baseline face image of each patient, or the baseline face image of each patient and the patient's identity information can be directly captured and counted in the central processing system, and then imported into the data processing system.

[0058] The image processing module is used to obtain the reference facial feature information of the reference facial image corresponding to each patient, and the patient's name, gender, age, ID number, etc. are recorded in the same list.

[0059] When treating each patient, the video capture mode of the camera acquisition component is first used to obtain the patient's real-time facial image video stream, and the real-time facial image video stream is subjected to image frame splitting, image frame screening and image frame analysis to obtain the real-time feature data of the real-time facial image corresponding to the patient. The comparison and recognition module is used to compare the current patient's real-time feature data with the baseline facial feature information of all patients to screen out the patient's identity information with matching features.

[0060] The camera acquisition component is then used to scan the QR code on the treatment sheet, and the identity information contained in the QR code on the treatment sheet is identified through the image processing module.

[0061] The comparison and recognition module is used to compare the identity information corresponding to the current patient with the identity information contained in the QR code to match the drug and the patient. If the comparison results are the same, subsequent treatment operations can be performed. If the comparison results are different, a second check is required.

[0062] It should be noted that since the application scenario of this system is mainly used by nurses to deliver medicines, the number of baseline facial image samples corresponding to all patients stored in the system when in use is relatively small. Therefore, the number of facial features used for face recognition can be relatively small, and the use of fuzzy algorithm for face matching can shorten the recognition time.

[0063] This embodiment adopts a video shooting mode, which can automatically capture instant facial image video streams in a short period of time. When the image processing module pre-processes the instant facial image video stream actually obtained by the camera acquisition component, it first splits the instant facial image video stream into multiple frames of facial images by performing image frame splitting and image frame screening on multiple frames of images in sequence, and uses the clear image of each frame of facial image as the first screening condition, and the expression state of the screened facial image as the second screening condition to select the real-time facial image for identifying the instant feature data of the current patient.

[0064] Specifically, by using the expression state of the facial image as the second screening condition, a real-time facial image that matches the expression state of the reference facial image can be automatically selected, and the number of the real-time facial images is ≥2, so that by performing subsequent image frame analysis on more than two real-time facial images, the real-time feature data on the real-time facial image is obtained, and the real-time feature data on multiple real-time facial images are secondary integrated to improve the accuracy of the acquired facial features, thereby improving the efficiency of subsequent facial feature matching, and fuzzy algorithms can be used for facial matching to reduce the difficulty of analysis in facial recognition programs.

[0065] Based on the above medical face recognition and treatment matching system, such as Figure 2 As shown, the present invention also provides a medical face recognition and treatment matching method, comprising the following steps:

[0066] Obtaining a baseline facial image of each patient, using a facial feature recognition algorithm to obtain the patient's baseline facial feature information, and saving the baseline facial feature information and original identity information of each patient;

[0067] A plurality of frames of images are intercepted from a video stream of a patient currently undergoing treatment to obtain a real-time facial image of the patient currently undergoing treatment, and a facial feature recognition algorithm is used to obtain facial features in each image of the real-time facial image, and the facial features of all images are integrated to obtain instant feature data of the patient currently undergoing treatment, and the original identity information of the patient currently undergoing treatment is obtained based on a comparison result between the instant feature data and the reference facial feature information;

[0068] Obtain the QR code information on the treatment form and identify the second identity information contained in the QR code; compare the original identity information and the second identity information, and only perform treatment when the information is completely matched. Preferably, extract the original identity information of the patient currently being treated as the medical record number (the ID card number may also be added), identify the second identity information contained in the QR code as the medical record number (the ID card number may also be added), compare the medical record number and the ID card number of the original identity information and the second identity information in sequence, and issue a voice prompt when the medical record numbers of the original identity information and the second identity information are exactly the same.

[0069] Before the specific treatment is carried out, the present embodiment uses the camera and the image processing module to adopt the face recognition method to automatically obtain the identity information of the current patient, and then uses the camera and the image processing module to identify the identity information contained in the QR code. The treatment is carried out only when the information is completely matched. This can ensure the correspondence between the medicine and the patient when the nurse performs the treatment, and prevent errors in the treatment.

[0070] In addition, when the present embodiment utilizes facial recognition to determine the patient's identity information, it specifically collects the facial video stream of the current patient, processes all frame images of the facial video stream in sequence to obtain a real-time facial image with high clarity and consistent with the expression state of the reference facial image, obtains the instant feature data of each image in the real-time facial image in sequence, and finally integrates all the instant feature data after error correction to obtain the instant feature data of the patient currently undergoing treatment. That is, before facial feature matching, the accuracy of the instant feature data of the patient currently undergoing treatment is improved by pre-processing the acquired facial image of the current patient, thereby reducing the feature requirements for facial feature matching. Using a fuzzy algorithm for face matching can shorten the recognition time and improve the recognition accuracy.

[0071] It should be noted that the facial feature recognition algorithm used in this embodiment adopts face detection fuzzy technology, specifically applying the Haar Cascade algorithm, which is a machine learning method for detecting objects such as faces in images or videos. It operates by utilizing a set of training features called "Haar-like features", which are simple rectangular filters that focus on changes in pixel intensity within an image area and can capture edges, angles and other features commonly found in faces. After training, the Haar Cascade essentially becomes a hierarchy of classifiers, with each stage gradually improving the detection process.

[0072] This embodiment can process the video stream of the current patient acquired in real time so that the number of facial features used for facial recognition is relatively small, and a relatively rough comparison method can be used to achieve a relatively high recognition accuracy.

[0073] Furthermore, the specific implementation method of obtaining the patient's baseline facial feature information is as follows:

[0074] Performing image processing on the reference face image to extract the facial features and facial curves of the patient, taking the center position of the patient's nostril as the central base point, and constructing a two-dimensional coordinate system based on the central base point;

[0075] Feature base points are selected from facial features and facial curves to form a facial base point polyline graph, and reference facial feature information is calculated respectively, wherein the reference facial feature information is divided into a left face facial feature group and a right face facial feature group.

[0076] As a preferred embodiment of the present embodiment, the facial recognition method in the present embodiment can still realize facial recognition even when the patient is sleeping, mainly relying on the following two aspects. On the one hand, when used by inpatient nurses, the number of facial feature samples of all patients stored is relatively small. Therefore, even if half-face facial feature matching is performed, the patient's identity can be identified smoothly and treatment matching can be performed. On the other hand, the present embodiment divides the baseline facial feature information into a left-face facial feature group and a right-face facial feature group. Therefore, when the patient lies on his side and all facial features of half of the face cannot be fully obtained, the patient's identity is identified by matching all features of the other half of the face.

[0077] Among them, the characteristic elements included in the baseline facial feature information are: the distance between the central base point and the nose-eye line between the front ends of the two eyes, the angle between the two nose-eye lines, the distance between the two sides of the nose wing and the central base point, the distance between the corners of the mouth on both sides, the angle between the corners of the mouth on both sides and the central base point, the upper and lower spacing of the curve where the mouth is located, and the distance between the central base point and each inflection point of the facial curve.

[0078] The above are only examples, but are not limited to the above feature elements. The feature elements used for facial recognition are determined specifically according to the facial feature recognition algorithm. The method of obtaining the above feature elements exemplified in this embodiment is simple. Since the number of basic samples is relatively small, most of them are patients in several wards, relatively simple and easy-to-obtain feature elements can be used for comparison, thereby reducing the total time consumed for facial recognition, improving facial recognition efficiency, and not hindering the overall facial recognition accuracy.

[0079] Furthermore, multiple frames of images are intercepted from the video stream of the patient currently undergoing treatment and a real-time facial image of the patient currently undergoing treatment is obtained. When the facial features in each image in the real-time facial image are obtained, multiple frames of images are intercepted from the video stream, and the multiple frames of images are first screened using clarity as a screening condition to screen out sequence facial images whose clarity is greater than a set threshold.

[0080] The facial state is used as a screening condition to perform secondary screening on the sequence facial images, so as to screen out the real-time facial images whose facial states meet the set expectations, take the real-time facial images as the processing objects, and obtain the instant feature data of the patient currently being treated from the real-time facial images;

[0081] The instant feature data of real-time facial images are integrated to form a set of multiple feature elements.

[0082] Furthermore, the implementation method of performing secondary screening on the sequence face images with the face state as the screening condition to screen out the real-time face images whose face state meets the set expectation is as follows:

[0083] Obtain a mouth curve formed by the mouth of each sequence of face images;

[0084] Taking the center base point of the two-dimensional coordinate system as the origin, multiple comparison base points are selected from the mouth curve, and the two-dimensional coordinate values ​​corresponding to the comparison base points are determined, and the comparison base points representing the corners of the mouth on both sides are determined based on the absolute values ​​of the X coordinates in all the two-dimensional coordinate values;

[0085] Determine a comparison base point representing the lowest position of the mouth based on the absolute value of the Y coordinate in all the two-dimensional coordinate values;

[0086] Calculate the facial state angle between the selected comparison base points representing the corners of the mouth on both sides and the comparison base point representing the lowest position of the mouth;

[0087] Select the sequence of face images whose facial state angles meet the set expectations as real-time face images.

[0088] In this embodiment, the specific advantage of using video streaming to obtain instant feature data of the real-time facial image of the current patient is that: according to the above, in this embodiment, a fuzzy algorithm is specifically used for facial feature matching, which reduces the difficulty of using the facial recognition and treatment matching system and improves convenience.

[0089] In order to improve the recognition accuracy of the fuzzy algorithm, it is necessary to screen the real-time face image of the current patient used for facial recognition, so as to ensure that the expression state of the real-time face image is consistent with the expression state of the reference face image, thereby improving the effectiveness of the instant feature data of the real-time face image.

[0090] For example, if the expression state of the reference face image is expressionless, then the expression state of the real-time face image selected from the video stream is also expressionless, thereby improving the accuracy of facial feature matching.

[0091] When the real-time facial image is processed to obtain the instant feature data of the patient currently undergoing treatment, the instant feature data is divided into a left face instant feature data group and a right face instant feature data group.

[0092] The same feature elements in the left face instant feature data group and the left face facial feature group of all benchmark facial images are compared respectively, and the same feature elements in the right face instant feature data group and the right face facial feature group of all benchmark facial images are compared respectively. The identity information of the current patient is determined by combining the left face comparison matching rate and the right face comparison matching rate.

[0093] Among them, when the left face comparison matching rate is greater than the set threshold, and the right face comparison matching rate is greater than the set threshold, it is considered that the real-time face image is the same as the reference face image, and the identity information of the current patient is determined.

[0094] When the left face comparison matching rate is less than a set threshold, or the right face comparison matching rate is less than a set threshold, it is considered that the real-time face image is different from the reference face image.

[0095] According to the above, the baseline facial features corresponding to the baseline facial image of each patient are divided into a left face facial feature group and a right face facial feature group. That is to say, the present application divides face recognition into left face recognition and right face recognition. When the matching rates of the left face and the right face both meet the set threshold, the identity information of the current patient is recognized and determined to improve the recognition accuracy.

[0096] In addition, when the face of the current patient is in a side-lying position, after obtaining the real-time feature data of the patient currently undergoing treatment, the number of feature elements in the left face real-time feature data group is compared with the number of feature elements in the right face real-time feature data group. When the number of feature elements in the left face real-time feature data group is not equal to the number of feature elements in the right face real-time feature data group, unilateral facial feature comparison is performed, and the set threshold of the left face comparison matching rate or the right face comparison matching rate is adjusted to increase. Therefore, when the patient is in a side-lying position, normal facial recognition can be performed to confirm the patient's identity and achieve matching between the patient and the drug.

[0097] On the other hand, the present invention also provides a medical face recognition and treatment matching device, such as Figure 3 and Figure 4 As shown, it includes: a PDA shell 3 and a PDA integrated system 4, the camera acquisition component 1 is installed on the head of the PDA shell 3, wherein the PDA integrated system 4 includes a data processing system 45, a battery module 43, and a touch screen display module 41 connected to the data processing system 45, a communication module 42, a voice broadcast module 44 and a data storage module 46;

[0098] A USB interface 2 is provided below the PDA housing 3, and a data processing system 45 is connected to the USB interface 2, and the USB interface 2 is used to import the reference facial images and identity information of all patients;

[0099] The data processing system 45 is used to receive information entered by the touch screen display module 41;

[0100] The data processing system 45 sends data to the host computer through the communication module 42;

[0101] The data processing system 45 controls the voice broadcast module 44 to broadcast information based on the comparison result between the patient identity information determined by face recognition and the patient information contained in the QR code on the treatment order.

[0102] The data processing system 3 processes the reference facial images of all patients in sequence to obtain the reference facial feature information of each patient, and transfers the reference facial images, reference facial feature information and identity information of all patients to the data storage module 46 .

[0103] When the nurse delivers medicine for treatment, the camera acquisition component 1 captures the video stream of the current patient, and the data processing system 3 compares the features of the screened real-time facial image of the current patient with the reference facial image to determine the identity information of the current patient. The data processing system 3 compares the QR code information collected by the camera acquisition component 1 with the identity information of the current patient. When the two are exactly the same, "Same identity" is played; when they are different, a "beep" alarm sound is played.

[0104] The touch screen display module 41 installed on the PDA housing 3 adopts a capacitive touch screen and can be operated by hand touch or electronic pen to enter patient images and identity information, output information, write data, and call the keyboard to facilitate users to input data.

[0105] The battery module 43 can provide stable power supply for the operation of each module component, and the battery module 43 can be charged through a power line.

[0106] The communication module 42 is used to send the data information of each patient corresponding to the PDA integrated system 4 to the host computer.

[0107] Furthermore, in this embodiment, the camera acquisition component 1 can be used to directly capture the patient's baseline facial image, and then the patient's baseline facial image and identity information can be imported into the host computer through the USB interface 2. The host computer can also acquire the baseline facial image of each patient, and then import the patient's baseline facial image and identity information into the PDA integrated system 4 through the USB interface 2.

[0108] The above embodiments are only exemplary embodiments of the present application and are not intended to limit the present application. The protection scope of the present application is defined by the claims. Those skilled in the art may make various modifications or equivalent substitutions to the present application within the essence and protection scope of the present application, and such modifications or equivalent substitutions shall also be deemed to fall within the protection scope of the present application.

Claims

1. A medical face recognition and treatment matching system, characterized in that: It includes a video acquisition component, a data processing system and a central processing system that is communicatively connected to the data processing system; The video acquisition component is used to obtain the instant facial image video stream of each patient and the QR code information on the treatment order; The data processing system includes a data storage module, an image processing module and a comparison and recognition module; The data storage module stores a reference face image of each patient, reference facial feature information of each patient acquired by the image processing module, and identity information of each patient; The image processing module sequentially performs image frame splitting, image frame screening and image frame analysis on the real-time face image video stream actually acquired by the camera acquisition component, and acquires the real-time feature data of the real-time face image of the current patient in real time, and the image processing module is used to perform image decoding and analysis on the two-dimensional code information on the treatment order to identify the identity information contained in the two-dimensional code; The comparison and recognition module compares the instant feature data of the real-time facial image of the current patient with the reference facial feature information of all patients to obtain the identity information corresponding to the current patient, and compares the identity information corresponding to the current patient with the identity information contained in the QR code to match the drug with the patient; The data processing system sends the reference facial image, identity information of each patient and the matching information of the medicine and the patient during each treatment to the central processing system for data storage.

2. A medical face recognition and treatment matching system according to claim 1, characterized in that: When the image processing module sequentially performs image frame splitting and image frame screening on the instant facial image video stream actually acquired by the camera acquisition component, the instant facial image video stream is first split into multiple frames of facial images, and the clear image of each frame of facial image is used as the first screening condition, and the expression state of the screened facial image is used as the second screening condition to select the real-time facial image for identifying the instant feature data of the current patient.

3. A medical face recognition and treatment matching method, characterized in that: A medical face recognition and treatment matching system according to any one of claims 1-2, comprising the following steps: Obtaining a baseline facial image of each patient, using a facial feature recognition algorithm to obtain the patient's baseline facial feature information, and saving the baseline facial feature information and original identity information of each patient; A plurality of frames of images are intercepted from a video stream of a patient currently undergoing treatment to obtain a real-time facial image of the patient currently undergoing treatment, and a facial feature recognition algorithm is used to obtain facial features in each image of the real-time facial image, and the facial features of all images are integrated to obtain instant feature data of the patient currently undergoing treatment, and the original identity information of the patient currently undergoing treatment is obtained based on a comparison result between the instant feature data and the reference facial feature information; Obtain the QR code information on the treatment order and identify the second identity information contained in the QR code; Compare the original identity information with the second identity information, and only proceed with treatment when the information matches completely.

4. A medical face recognition and treatment matching method according to claim 3, characterized in that: The specific implementation method of obtaining the patient's baseline facial feature information is as follows: Performing image processing on the reference face image to extract facial features and facial curves of the patient, taking the center position of the patient's nostril as a central base point, and constructing a two-dimensional coordinate system according to the central base point; Feature base points are selected from facial features and facial curves to form a facial base point polyline graph, and reference facial feature information is calculated respectively, wherein the reference facial feature information is divided into a left face facial feature group and a right face facial feature group.

5. A medical face recognition and treatment matching method according to claim 4, characterized in that: The characteristic elements included in the baseline facial feature information are: the distance between the central base point and the nose-eye line between the front ends of the two eyes, the angle between the two nose-eye lines, the distance between the two sides of the nose wing and the central base point, the distance between the corners of the mouth on both sides, the angle between the corners of the mouth on both sides and the central base point, the upper and lower spacing of the curve where the mouth is located, and the distance between the central base point and each inflection point of the facial curve.

6. A medical face recognition and treatment matching method according to claim 5, characterized in that: A plurality of frames of images are intercepted from the video stream, and the plurality of frames of images are first screened using definition as a screening condition to screen out sequence face images whose definition is greater than a set threshold; Performing secondary screening on the sequence of facial images with facial status as a screening condition to screen out real-time facial images whose facial status meets the set expectations, taking the real-time facial images as processing objects, and obtaining instant feature data of the patient currently undergoing treatment from the real-time facial images; The instant feature data of the real-time facial image are integrated to form a set consisting of multiple feature elements.

7. A medical face recognition and treatment matching method according to claim 6, characterized in that: The method for performing secondary screening on the sequence of face images with the face state as the screening condition to screen out the real-time face images whose face states meet the set expectations is as follows: Obtaining a mouth curve formed by the mouth of each of the sequence of facial images; Taking the center base point of the two-dimensional coordinate system as the origin, selecting multiple comparison base points from the mouth curve, and determining the two-dimensional coordinate values ​​corresponding to the comparison base points, and determining the comparison base points representing the corners of the mouth on both sides based on the absolute values ​​of the X coordinates in all the two-dimensional coordinate values; Determine a comparison base point representing the lowest position of the mouth based on the absolute value of the Y coordinate in all the two-dimensional coordinate values; Calculate the facial state angle between the selected comparison base points representing the corners of the mouth on both sides and the comparison base point representing the lowest position of the mouth; Select the sequence of face images whose facial state angles meet the set expectations as real-time face images.

8. A medical face recognition and treatment matching method according to claim 7, characterized in that: When the real-time facial image is processed to obtain the instant feature data of the patient currently undergoing treatment, the instant feature data is divided into a left face instant feature data group and a right face instant feature data group; Compare the same feature elements in the left face instant feature data group with the left face facial feature groups of all reference facial images, and compare the same feature elements in the right face instant feature data group with the right face facial feature groups of all reference facial images, and determine the identity information of the current patient by combining the left face comparison matching rate and the right face comparison matching rate; When the left face comparison matching rate is greater than the set threshold, and the right face comparison matching rate is greater than the set threshold, it is considered that the real-time face image is the same as the reference face image, and the identity information of the current patient is determined; When the left face comparison matching rate is less than a set threshold, or the right face comparison matching rate is less than a set threshold, it is considered that the real-time face image is different from the reference face image.

9. A medical face recognition and treatment matching method according to claim 8, characterized in that: After obtaining the real-time feature data of the patient currently undergoing treatment, the number of feature elements in the left face real-time feature data group is compared with the number of feature elements in the right face real-time feature data group. When the number of feature elements in the left face real-time feature data group is not equal to the number of feature elements in the right face real-time feature data group, unilateral facial feature comparison is performed, and the set threshold of the left face comparison matching rate or the right face comparison matching rate is adjusted to increase.

10. A medical face recognition and treatment matching method according to claim 3, characterized in that: The original identity information of the patient currently undergoing treatment is extracted as the medical record number, the second identity information contained in the QR code is identified as the medical record number, the medical record numbers of the original identity information and the second identity information are compared in sequence, and a voice prompt is issued when the medical record numbers of the original identity information and the second identity information are exactly the same.

Images