CN114429676A - Medical institution disinfection supply room personnel identity and behavior recognition system - Google Patents
Medical institution disinfection supply room personnel identity and behavior recognition system Download PDFInfo
- Publication number
- CN114429676A CN114429676A CN202210097699.2A CN202210097699A CN114429676A CN 114429676 A CN114429676 A CN 114429676A CN 202210097699 A CN202210097699 A CN 202210097699A CN 114429676 A CN114429676 A CN 114429676A
- Authority
- CN
- China
- Prior art keywords
- behavior
- supply room
- disinfection supply
- identity
- medical institution
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H40/00—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices
- G16H40/20—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices for the management or administration of healthcare resources or facilities, e.g. managing hospital staff or surgery rooms
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Biomedical Technology (AREA)
- Theoretical Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Business, Economics & Management (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Primary Health Care (AREA)
- Life Sciences & Earth Sciences (AREA)
- Medical Informatics (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Public Health (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Epidemiology (AREA)
- Medical Treatment And Welfare Office Work (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to a medical institution disinfection supply room personnel identity and behavior recognition system, which belongs to the technical field of image recognition and comprises: the system comprises a cloud platform, a video acquisition module, an identity recognition module, a behavior recognition module and a face database; the video acquisition module is arranged in the medical institution disinfection supply room and used for acquiring video signals in the medical institution disinfection supply room and submitting the video signals to the cloud platform; the identity recognition module is used for acquiring video signals from the cloud platform and performing living body identification and identity recognition on personnel entering a medical institution disinfection supply room according to the face database and the acquired video signals; the behavior recognition module is used for acquiring video signals from the cloud platform, and performing behavior classification recognition on personnel entering the medical institution disinfection supply room according to the acquired video signals based on the 3D convolutional neural network. The invention improves the accuracy of identity recognition and realizes the recognition of the behavior of the staff.
Description
Technical Field
The invention relates to the technical field of image recognition, in particular to a system for recognizing personnel identity and behavior in a disinfection supply room of a medical institution.
Background
Generally, the disinfection action of medical institutions on medical instruments occurs in a disinfection supply room, and the disinfection supply room has the characteristics of centralized disinfection places, uniform disinfection modes, less instruments and the like, and is easy for health supervision. Therefore, the disinfection supply room plays an important role in medical institution departments and is responsible for cleaning, disinfecting, packaging, storing and supplying all-mechanism medical instruments. In the field of health supervision, as a part of the indispensable safety guarantee of medical institutions, the normative operation of workers in a disinfection supply room is an important link for reducing the threat of bacteria, because the iatrogenic infection crisis is easy to occur once the situations such as improper operation occur.
Most of face recognition methods adopted by the existing health supervision system for the medical institution disinfection supply room only simply extract the characteristic information of a shot face image, and the image is not preprocessed to improve the face recognition precision and the recognition success rate and reduce the equipment memory consumption, so that the equipment cost is increased and the face recognition time is longer; except that the human face image is not preprocessed, most of the existing health supervision systems are not provided with a living body identification function, and the personnel entering a disinfection supply room are not subjected to accurate identity identification so as to generate potential safety hazards; moreover, most health supervision systems do not pay attention and conduct monitoring to whether the operation of workers in the disinfection supply room is carried out seriously or not and whether the operation meets the standards, so that the problems of safety omission exist, and even serious results can be caused.
Disclosure of Invention
The invention aims to provide a personnel identity and behavior identification system for a disinfection supply room of a medical institution, which improves the accuracy of identity identification and realizes the identification of the behavior of workers.
In order to achieve the purpose, the invention provides the following scheme:
a medical facility disinfection supply room personnel identity and behavior identification system comprising: the system comprises a cloud platform, a video acquisition module, an identity recognition module, a behavior recognition module and a face database;
the video acquisition module is arranged in a medical institution disinfection supply room and used for acquiring video signals in the medical institution disinfection supply room and submitting the video signals to the cloud platform;
the identity recognition module is used for acquiring video signals from the cloud platform and performing living body identification and identity recognition on personnel entering the medical institution disinfection supply room according to the human face database and the acquired video signals;
the behavior recognition module is used for acquiring video signals from the cloud platform, and performing behavior classification recognition on personnel entering the medical institution disinfection supply room according to the acquired video signals based on a 3D convolutional neural network.
Optionally, the identity module includes:
the face capturing unit is used for acquiring video signals from the cloud platform every first set unit time, decomposing the video signals in every first set unit time frame by frame to obtain frame images, and extracting the features of the frame images by adopting a convolutional neural network to acquire face feature images;
the living body identification unit is used for judging whether a movable person corresponding to the human face characteristic image exists in the medical institution disinfection supply room or not according to the human face characteristic image continuously obtained within set time;
the human face image preprocessing unit is used for carrying out gray level adjustment and image size normalization on the acquired human face characteristic image to obtain a preprocessed human face characteristic image when the living body identification unit judges that a movable person corresponding to the human face characteristic image exists in the medical institution disinfection supply room;
the face feature extraction unit is used for extracting the features of the preprocessed face feature image through a convolutional neural network to obtain a face image feature value;
and the face recognition unit is used for comparing the face image characteristic value with the face characteristic value in the face database to obtain the identity of the movable personnel.
Optionally, the first set unit time is 5 seconds.
Optionally, the face features in the face feature image include human head, eyes, nose and mouth.
Optionally, the living body identification unit includes a living body identification subunit, and the living body identification subunit is configured to determine, according to the face feature images continuously obtained within a set time, whether there is a behavior of blinking or a change of lips in the medical institution disinfection supply room, determine that there is a moving person corresponding to the face feature image in the medical institution disinfection supply room if yes, and determine that there is no moving person corresponding to the face feature image in the medical institution disinfection supply room if no.
Optionally, the behavior recognition module includes:
the video characteristic extraction unit is used for acquiring video signals from the cloud platform every second set unit time, extracting a characteristic graph from the video signals in every second set unit time through a 3D (three-dimensional) convolutional neural network, and recording the characteristic graph as a video information characteristic graph;
a candidate time sequence segment generating unit, configured to perform 3D convolution and 3D maximum pooling on the video information feature map in each second set unit time, and extract a human behavior time sequence segment and an environment time sequence segment;
and the behavior classification and identification unit is used for extracting fixed size features from the human behavior time sequence segment and the environment time sequence segment by adopting 3D Rol operation, and performing action classification and boundary regression on the human behavior time sequence segment and the environment time sequence segment in the extracted fixed size features based on feature integration to identify human behaviors.
Optionally, the action classification includes canonical behavior and abnormal behavior;
the normative behavior comprises that a person wears gloves when cleaning the medical instrument; the abnormal behavior includes cleaning the medical device without wearing gloves.
Optionally, the second set unit time is 30 seconds.
Optionally, the method further comprises: and the result display module is used for displaying the result of the identity recognition module and displaying the behavior classification recognition result of the behavior recognition module.
Optionally, the video capture module comprises a camera mounted within the medical facility disinfection supply room.
According to the specific embodiment provided by the invention, the invention discloses the following technical effects:
the invention determines that the medical institution disinfection supply room is a moving person based on the living body identification unit, compares the human face characteristic image preprocessed by the moving person with the human face characteristic value in the human face database through characteristic extraction to obtain the identity of the moving person, thereby improving the accuracy of identity identification.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings without inventive exercise.
FIG. 1 is a first schematic structural diagram of a medical facility disinfection supply room personnel identity and behavior recognition system according to the present invention;
FIG. 2 is a schematic structural diagram of a system for identifying personnel identity and behavior in a disinfection supply room of a medical institution according to the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The invention aims to provide a personnel identity and behavior identification system for a disinfection supply room of a medical institution, which improves the accuracy of identity identification and realizes the identification of the behavior of workers.
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in further detail below.
FIG. 1 is a first schematic structural diagram of a medical facility disinfection supply room personnel identity and behavior recognition system according to the present invention; FIG. 2 is a schematic diagram of a second embodiment of a system for identifying identity and behavior of personnel in a disinfection supply room of a medical institution according to the present invention; as shown in fig. 1-2, a medical facility disinfection supply room personnel identity and behavior identification system comprises: the system comprises a cloud platform 102 (unified monitoring cloud platform), a video acquisition module 101, an identity recognition module 103, a behavior recognition module 104 and a face database 105.
The video acquisition module 101 is disposed in the medical institution disinfection supply room, and is configured to acquire a video signal in the medical institution disinfection supply room and submit the video signal to the cloud platform 102.
The video acquisition module 101 comprises a high-definition camera installed in a disinfection supply room of a medical institution, acquires a video signal of a monitoring picture through the high-definition camera, and transmits the video signal back to the unified monitoring cloud platform 102 by using a network.
The cloud platform 102 is configured to receive, store, and parse the video signal from the video capture module 101.
The video signal storage means receives and stores the video signal transmitted from the video acquisition module 101;
the video signal analysis means that the video signals stored in the video signal storage module are respectively sent to the identity recognition module 103 and the behavior recognition module 104, and the video signal analysis comprises two parts: 1. sending the video signal to the identification module 103 in 5 seconds as a time unit; 2. the video signal is transmitted to the behavior recognizing module 104 in one time unit of 30 seconds.
The identification module 103 is used for acquiring video signals from the cloud platform 102 and performing living body identification and identification on the personnel entering the medical institution disinfection supply room according to the human face database 105 and the acquired video signals.
The identity recognition module 103 acquires a video signal of one unit of 5 seconds from the unified monitoring cloud platform, firstly decomposes the video signal into a plurality of pieces of image information frame by frame, invokes a face capturing function, performs feature analysis on the image information one by one, judges whether face features are included, and if the face features are included, performs a series of work of living body identification, face image preprocessing, face feature extraction and face recognition on the face image information in sequence to complete identity recognition.
The behavior recognition module 104 is configured to acquire a video signal from the cloud platform 102, and perform behavior classification recognition on the person entering the medical institution disinfection supply room according to the acquired video signal based on the 3D convolutional neural network, so as to determine whether the disinfection behavior is normal.
The identification module 103 includes:
the face capturing unit is used for acquiring video signals from the cloud platform 102 every first set unit time, decomposing the video signals in every first set unit time frame by frame to obtain frame images, and extracting features of the frame images by using a convolutional neural network to acquire face feature images.
The face features in the face feature image comprise the head, eyes, nose and mouth of a human body.
The face capturing unit is configured to perform feature extraction on an image, which is obtained by frame-by-frame decomposition of a video signal with a unit of 5 seconds acquired from the unified monitoring cloud platform, by using a Convolutional Neural Network (CNN), and if it is detected that the features include information such as a human head, eyes, and a mouth, it is determined that the image includes face information, the face feature image is extracted, and the identity recognition module 103 continues to perform other recognition operations on the face feature image.
And the living body identification unit is used for judging whether movable personnel corresponding to the human face characteristic image exists in the medical institution disinfection supply room or not according to the human face characteristic image continuously obtained within the set time.
The living body identification unit carries out characteristic judgment aiming at behaviors such as blinking and lip change and the like according to all the human face image characteristics calculated by the face capture unit, if the behavior of blinking or lip change is detected, the human face is judged to be from a real living body instead of a static picture, the accuracy of the human face image is further improved, the safety of a disinfection supply room is improved, and the identified human face image is used by the human face image preprocessing module.
The living body identification unit comprises a living body identification subunit, the living body identification subunit is used for judging whether the behavior of blinking or lip change exists in the medical institution disinfection supply room according to the face characteristic images continuously obtained within the set time, if so, the living body identification subunit judges that the moving personnel corresponding to the face characteristic images exist in the medical institution disinfection supply room, and if not, the living body identification subunit judges that the moving personnel corresponding to the face characteristic images do not exist in the medical institution disinfection supply room.
And the human face image preprocessing unit is used for carrying out gray level adjustment and image size normalization on the acquired human face characteristic image to obtain a preprocessed human face characteristic image when the living body identification unit judges that a movable person corresponding to the human face characteristic image exists in the medical institution disinfection supply room.
The human face image preprocessing unit firstly performs gray level adjustment on the human face image passing through the living body identification unit by means of linear gray level adjustment, power transformation, logarithm transformation and the like, so that the interference of light and shadow on image information is reduced, and human face expression is enhanced; and then, normalizing the size of the face feature image by using the image interpolation value to ensure that the input formats of the face feature extraction modules in the next step are consistent.
And the face feature extraction unit is used for extracting the features of the preprocessed face feature image through a convolutional neural network to obtain a face image feature value.
And the face recognition unit is used for comparing the face image characteristic value with the face characteristic value in the face database 105 to obtain the identity of the moving person.
The face feature extraction unit and the face recognition unit are combined with the face database 105 to realize identity recognition, the preprocessed face feature image is input into a CNN convolutional neural network, the final feature value of the face feature image is obtained through calculation of a convolutional layer, a pooling layer and a full connection layer, similarity comparison is carried out on the feature value and the face feature in the face database 105, identity information is determined, and identity recognition is completed.
The behavior recognition module 104 includes:
and the video feature extraction unit is used for acquiring the video signal from the cloud platform 102 every second set unit time, and extracting a feature map from the video signal in every second set unit time through a 3D convolutional neural network, and recording the feature map as a video information feature map.
And the candidate time sequence segment generating unit is used for performing 3D convolution and 3D maximum pooling operation on the video information characteristic graph in every second set unit time and extracting a human behavior time sequence segment and an environment time sequence segment.
And the behavior classification and identification unit is used for extracting fixed size characteristics from the human behavior time sequence segments and the environment time sequence segments by adopting 3D Rol operation, and performing action classification and boundary regression on the human behavior time sequence segments and the environment time sequence segments in the extracted fixed size characteristics based on feature integration to identify human behaviors.
The action classification comprises a normative action and an abnormal action;
normative behaviors include wearing gloves by personnel when cleaning medical instruments; the abnormal behavior includes cleaning the medical device without wearing gloves.
The second set unit time is 30 seconds.
The behavior recognition module 104 is used for capturing the motion category and locating the start time and the end time of the motion, and is composed of a video feature extraction unit, a candidate time sequence segment generation unit and a behavior classification recognition unit. After the behavior recognition module 104 acquires a video signal with a unit of 30 seconds from the unified monitoring cloud platform, firstly, original video information is converted into 3-dimensional feature information through a 3D convolutional neural network in a video feature extraction unit, and a feature map is extracted; secondly, performing 3D convolution and 3D maximum pooling operation on the feature map extracted by the video feature extraction unit by using a candidate time sequence fragment module to extract a required human behavior time sequence fragment and an environment time sequence fragment; and finally, mining fixed-size characteristics of partial human body behavior time sequence segments and environmental time sequence segments by using 3D Rol operation in the behavior classification and identification unit to enable the fixed-size characteristics to have the same dimensionality and be connected with a full-connection layer, and then performing action classification (standard behavior and abnormal behavior) and boundary regression (adjusting the center and the length of an action video segment) on the selected partial human body behavior time sequence segments and environmental time sequence segments on the basis of feature integration to realize human body behavior identification. The candidate time sequence segment generation unit and the behavior classification and identification unit share the feature map extracted by the 3D convolutional neural network in the video feature extraction unit.
The normative behaviors refer to standard technical operation rules for cleaning, disinfecting, storing and the like of medical instruments specified by a supply room of a disinfection room of a medical institution, and require that working personnel must strictly execute the operations, otherwise serious consequences such as iatrogenic infection and the like can be caused. Wherein the normative behaviors are that when workers clean medical instruments conventionally, the workers need to wear medical gloves to prevent injury and infection; the corresponding abnormal behavior means that the staff does not wear medical gloves to directly clean the medical articles when cleaning the medical articles. Such a case is more of normative behavior and abnormal behavior.
The specific process of identifying whether the cleaning operation is standardized by the behavior identification module 104 is as follows: (1) after the behavior recognition module 104 acquires a video signal with 30 seconds as a unit from the unified monitoring cloud platform, original video information is converted into 3-dimensional feature information through a 3D convolutional neural network in a video feature extraction unit, and a feature map in the 30-second video is extracted. (2) And performing 3D convolution and 3D maximum pooling operation on the feature map extracted by the video feature extraction unit by using the candidate time sequence segment unit to extract a human behavior time sequence segment and an environment time sequence segment which are required for judging whether the cleaning operation is standard or not. Assuming that the 4 th second picture is extracted to be in the center of the cleaning pool, the two hands of the person wear medical gloves to hold medical instruments; the picture of the 5 th second is that one hand of a person wears the medical gloves to hold medical instruments and the other hand wears the medical gloves to contact a faucet switch in the center of the cleaning pool; the picture of the 6 th second is that the tap switch is displaced outwards when the hands of a person wear medical gloves to hold medical instruments in the center of the cleaning pool; the picture of the 7 th second is that in the center of the cleaning pool, one hand of the person wears the medical gloves to hold the medical instruments, and the other hand wears the medical gloves to hold the cleaning brushes; the picture of the 10 th second is that in the center of the cleaning pool, one hand of a person wears medical gloves to hold medical instruments, and the other hand wears medical gloves to hold a cleaning brush to contact with the disinfection cleaning solution; the 21 st frame is the frame in the center of the washing tank, in which the user wears medical gloves with his hands and holds the medical instrument, and the faucet is turned back to the home position. (3) And mining partial human body behavior time sequence segments and environmental time sequence segments by using 3D Rol operation in the behavior classification and identification unit, and judging that the cleaning operation meets the specification.
The specific process of the behavior recognition module 104 for recognizing whether the disinfection operation is standard is as follows: (1) after the behavior recognition module 104 acquires a video signal with 30 seconds as a unit from the unified monitoring cloud platform, original video information is converted into 3-dimensional feature information through a 3D convolutional neural network in a video feature extraction unit, and a feature map in the 30-second video is extracted. (2) And performing 3D convolution and 3D maximum pooling operation on the feature map extracted by the video feature extraction unit by using the candidate time sequence segment module to extract a human behavior time sequence segment and an environment time sequence segment which are required for judging whether the disinfection operation is standard or not. If the extracted picture of the 3 rd second is in a working state of the sterilizer, namely the screen of the sterilizer is dark; the picture of the 5 th second is that one hand of the person wears medical gloves to contact with the sterilizer; the 6 th picture is the opening state of the sterilizer door; the 10 th second frame is a sterilizer screen highlight state. (3) And mining partial human body behavior time sequence segments and environmental time sequence segments by using 3D Rol operation in the behavior classification and identification unit, and judging that the disinfection operation meets the specification.
The process of identifying whether the packaging and storage operations are standardized by the behavior identification module 104 is similar to the above, and whether the operations meet the standard can be determined by identifying the key behavior pictures.
Specifically, the video feature extraction unit is used for convolving three dimensions of length, width and height of a segmented video sequence frame which is acquired from a unified monitoring cloud platform and takes 30 seconds as a unit by using a 3D convolutional neural network, calculating a feature value of the segmented video sequence frame, and extracting video features.
The candidate time sequence segment generating unit extracts video frame segments containing human body behaviors in the video stream. Step one, receiving a feature map extracted by a video feature extraction unit as input, generating K candidate time sequence anchors with different lengths at each position of a time domain with uniformly distributed features on the feature map, step two, expanding a time sequence perception range through 3D convolution in order to obtain the features of each time position, then predicting and grading all anchors by using 3D maximum pooling, step three, extracting a human behavior time sequence segment with the first 30% of the score value and an environment time sequence segment with the last 30% of the score value, and discarding the others.
The behavior classification and identification unit selects partial human behavior time sequence segments and environmental time sequence segments to perform 3DRol pooling operation, converts feature maps with any size into feature maps with the same size, is convenient to input into a full-connection structure, then inputs the features of the selected partial human behavior time sequence segments and environmental time sequence segments into a full-connection layer, classifies (normative behaviors and abnormal behaviors) through the classification layer, adjusts the starting time and the ending time of the actions through the regression layer, and realizes human behavior identification through training.
The face database 105 is used for collecting full face information images in multiple directions for formal staff in disinfection supply rooms of all medical institutions in a set range (such as a certain city), and importing the collected image information marked with feature tags into the database to serve as a standard comparison face database 105.
A medical facility disinfection supply room personnel identity and action identification system further comprising: and the result display module 106 is configured to display the result of the identity recognition performed by the identity recognition module 103, and display the result of the behavior classification recognition performed by the behavior recognition module 104.
The result display module 106 includes an analysis database and a graphical report unit, and is mainly used for statistics and display of results.
The analysis database is a database for storing analysis results of the identity recognition module 103 and the behavior recognition module 104, such as face matching results and whether abnormal results exist in behaviors, and has a statistical function for inquiring at any time and providing data support for intelligent health supervision.
The graphical report unit shows the results stored in the analysis database in the forms of graphs and the like, so that readability and readability are enhanced, comprehensibility is facilitated, communication efficiency is improved, and intelligent health supervision decisions are assisted.
The personnel identity and behavior identification system for the disinfection supply room of the medical institution monitors the whole disinfection process in real time and can effectively monitor the omnibearing safety problem of the disinfection supply room.
Firstly, the living body identification function is set, the problem that safety threat is caused when formal staff in a disinfection supply room of a non-medical institution intentionally enters the disinfection supply room is mainly monitored, the missing part of a general face recognition step is made up, the safety problem of the disinfection supply room of the medical institution is completely guaranteed, the supervision flow of the disinfection supply room is optimized, the safety is improved, additional monitoring equipment is not needed, and the equipment and the cost of the medical institution are reduced.
Secondly, the invention sets an image preprocessing function and is internally provided with a plurality of high-efficiency image preprocessing algorithms, thereby reducing the influence of external factors such as light, noise, size, contour and the like on the quality of the acquired face image and improving the success rate and the accuracy rate of face image recognition.
Finally, the invention adopts a 3D convolutional neural network (3-dimensional convolutional neural networks) to identify whether the behaviors of workers in the disinfection supply room are standardized, thereby improving the identification efficiency, reducing the labor cost, realizing automation and intellectualization, taking the obtained analysis data as the basis whether the medical institution strictly follows the standardized disinfection flow, providing data support for health supervision, improving the quality of medical care, and preventing and controlling the occurrence of the iatrogenic infection problem of the medical institution to a greater extent, and having important significance.
The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other.
The principles and embodiments of the present invention have been described herein using specific examples, which are provided only to help understand the method and the core concept of the present invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, the specific embodiments and the application range may be changed. In view of the above, the present disclosure should not be construed as limiting the invention.
Claims (10)
1. A medical facility disinfection supply room personnel identity and behavior recognition system, comprising: the system comprises a cloud platform, a video acquisition module, an identity recognition module, a behavior recognition module and a face database;
the video acquisition module is arranged in a medical institution disinfection supply room and is used for acquiring video signals in the medical institution disinfection supply room and submitting the video signals to the cloud platform;
the identity recognition module is used for acquiring video signals from the cloud platform and performing living body identification and identity recognition on personnel entering the medical institution disinfection supply room according to the human face database and the acquired video signals;
the behavior recognition module is used for acquiring video signals from the cloud platform, and performing behavior classification recognition on personnel entering the medical institution disinfection supply room according to the acquired video signals based on a 3D convolutional neural network.
2. The medical facility disinfection supply room personnel identity and action identification system of claim 1, wherein said identity module comprises:
the face capturing unit is used for acquiring video signals from the cloud platform every first set unit time, decomposing the video signals in every first set unit time frame by frame to obtain frame images, and extracting the features of the frame images by adopting a convolutional neural network to acquire face feature images;
the living body identification unit is used for judging whether a movable person corresponding to the human face characteristic image exists in the medical institution disinfection supply room or not according to the human face characteristic image continuously obtained within set time;
the human face image preprocessing unit is used for carrying out gray level adjustment and image size normalization on the acquired human face characteristic image to obtain a preprocessed human face characteristic image when the living body identification unit judges that a movable person corresponding to the human face characteristic image exists in the medical institution disinfection supply room;
the face feature extraction unit is used for extracting the features of the preprocessed face feature image through a convolutional neural network to obtain a face image feature value;
and the face recognition unit is used for comparing the face image characteristic value with the face characteristic value in the face database to obtain the identity of the movable personnel.
3. The medical facility disinfection supply room personnel identity and action identification system of claim 2 wherein said first set unit time is 5 seconds.
4. The medical institution disinfection supply room personnel identity and behavior recognition system of claim 2, wherein the facial features in the facial feature image include a human head, eyes, nose and mouth.
5. The system of claim 2, wherein the living body identification unit comprises a living body identification subunit, and the living body identification subunit is configured to determine whether there is blinking behavior or lip changing behavior in the disinfection supply room of the medical institution according to the facial feature images continuously obtained within a set time, determine that there is a moving person corresponding to the facial feature image in the disinfection supply room of the medical institution if the blinking behavior or lip changing behavior is determined, and determine that there is no moving person corresponding to the facial feature image in the disinfection supply room of the medical institution if the blinking behavior or lip changing behavior is determined.
6. The medical facility disinfection supply room personnel identity and behavior identification system of claim 1, wherein said behavior identification module comprises:
the video characteristic extraction unit is used for acquiring video signals from the cloud platform every second set unit time, extracting a characteristic graph from the video signals in every second set unit time through a 3D (three-dimensional) convolutional neural network, and recording the characteristic graph as a video information characteristic graph;
a candidate time sequence segment generating unit, configured to perform 3D convolution and 3D maximum pooling on the video information feature map in each second set unit time, and extract a human behavior time sequence segment and an environment time sequence segment;
and the behavior classification and identification unit is used for extracting fixed size features from the human behavior time sequence segment and the environment time sequence segment by adopting 3D Rol operation, and performing action classification and boundary regression on the human behavior time sequence segment and the environment time sequence segment in the extracted fixed size features based on feature integration to identify human behaviors.
7. The medical facility disinfection supply room personnel identity and behavior identification system of claim 6, wherein said action classification includes normative behavior and abnormal behavior;
the normative behavior comprises that a person wears gloves when cleaning the medical instrument; the abnormal behavior includes cleaning the medical device without wearing gloves.
8. The medical facility disinfection supply room personnel identity and action identification system of claim 6 wherein said second predetermined unit of time is 30 seconds.
9. The medical facility disinfection supply room personnel identity and action identification system of claim 1, further comprising: and the result display module is used for displaying the result of the identity recognition module and displaying the behavior classification recognition result of the behavior recognition module.
10. The medical facility disinfection supply room personnel identity and action identification system of claim 1, wherein said video capture module comprises a camera mounted within said medical facility disinfection supply room.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210097699.2A CN114429676B (en) | 2022-01-27 | 2022-01-27 | Personnel identity and behavior recognition system for disinfection supply room of medical institution |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210097699.2A CN114429676B (en) | 2022-01-27 | 2022-01-27 | Personnel identity and behavior recognition system for disinfection supply room of medical institution |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114429676A true CN114429676A (en) | 2022-05-03 |
| CN114429676B CN114429676B (en) | 2023-07-25 |
Family
ID=81312397
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210097699.2A Active CN114429676B (en) | 2022-01-27 | 2022-01-27 | Personnel identity and behavior recognition system for disinfection supply room of medical institution |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114429676B (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101770613A (en) * | 2010-01-19 | 2010-07-07 | 北京智慧眼科技发展有限公司 | Social insurance identity authentication method based on face recognition and living body detection |
| CN111291699A (en) * | 2020-02-19 | 2020-06-16 | 山东大学 | Substation personnel behavior identification method based on monitoring video time sequence action positioning and abnormity detection |
| CN111339126A (en) * | 2020-02-27 | 2020-06-26 | 平安医疗健康管理股份有限公司 | Medical data screening method and device, computer equipment and storage medium |
| CN111695542A (en) * | 2020-06-23 | 2020-09-22 | 韶关市启之信息技术有限公司 | Video monitoring fine-grained analysis method and system |
| CN112183390A (en) * | 2020-09-30 | 2021-01-05 | 中国人民解放军陆军军医大学第一附属医院 | Medical personnel hand washing monitoring system |
| CN112397183A (en) * | 2020-10-12 | 2021-02-23 | 深圳市豫知科技有限公司 | Operating room monitoring and management system based on biological information identification |
| CN112507830A (en) * | 2020-11-30 | 2021-03-16 | 京东方科技集团股份有限公司 | High-risk area protection management and control method, system and storage medium |
| CN113765750A (en) * | 2021-09-07 | 2021-12-07 | 山东纬横数据科技有限公司 | Non-invasive system survival monitoring system |
-
2022
- 2022-01-27 CN CN202210097699.2A patent/CN114429676B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101770613A (en) * | 2010-01-19 | 2010-07-07 | 北京智慧眼科技发展有限公司 | Social insurance identity authentication method based on face recognition and living body detection |
| CN111291699A (en) * | 2020-02-19 | 2020-06-16 | 山东大学 | Substation personnel behavior identification method based on monitoring video time sequence action positioning and abnormity detection |
| CN111339126A (en) * | 2020-02-27 | 2020-06-26 | 平安医疗健康管理股份有限公司 | Medical data screening method and device, computer equipment and storage medium |
| CN111695542A (en) * | 2020-06-23 | 2020-09-22 | 韶关市启之信息技术有限公司 | Video monitoring fine-grained analysis method and system |
| CN112183390A (en) * | 2020-09-30 | 2021-01-05 | 中国人民解放军陆军军医大学第一附属医院 | Medical personnel hand washing monitoring system |
| CN112397183A (en) * | 2020-10-12 | 2021-02-23 | 深圳市豫知科技有限公司 | Operating room monitoring and management system based on biological information identification |
| CN112507830A (en) * | 2020-11-30 | 2021-03-16 | 京东方科技集团股份有限公司 | High-risk area protection management and control method, system and storage medium |
| CN113765750A (en) * | 2021-09-07 | 2021-12-07 | 山东纬横数据科技有限公司 | Non-invasive system survival monitoring system |
Non-Patent Citations (1)
| Title |
|---|
| 周世乐;莫景书;: "消毒供应质量追溯管理系统对外来医疗器械全流程监控管理的效果分析", 护士进修杂志, no. 10, pages 898 - 900 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114429676B (en) | 2023-07-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110826538B (en) | Abnormal off-duty identification system for electric power business hall | |
| Yu et al. | An online one class support vector machine-based person-specific fall detection system for monitoring an elderly individual in a room environment | |
| CN110458101B (en) | Criminal personnel sign monitoring method and equipment based on combination of video and equipment | |
| Hu et al. | Research on abnormal behavior detection of online examination based on image information | |
| CN110717542A (en) | Emotion recognition method, device and equipment | |
| CN110264668A (en) | More tactful old men based on machine vision technique see maintaining method | |
| CN112528952B (en) | Working state intelligent recognition system for electric power business hall personnel | |
| CN111476114A (en) | Fatigue detection method, device, terminal equipment and storage medium | |
| CN112768070A (en) | Mental health evaluation method and system based on dialogue communication | |
| CN115205764B (en) | Online learning concentration monitoring method, system and medium based on machine vision | |
| CN113920461A (en) | Power grid operation and maintenance process image monitoring system and monitoring method | |
| CN112287741A (en) | Image processing-based farming operation management method and device | |
| CN113591550B (en) | Method, device, equipment and medium for constructing personal preference automatic detection model | |
| AU2021218192B1 (en) | Video anomaly detection method and system based on generative assistant discriminative network | |
| CN113314230A (en) | Intelligent epidemic prevention method, device, equipment and storage medium based on big data | |
| CN114429676A (en) | Medical institution disinfection supply room personnel identity and behavior recognition system | |
| CN112488647A (en) | Attendance system and method, storage medium and electronic equipment | |
| CN110852284A (en) | System for predicting user concentration degree based on virtual reality environment and implementation method | |
| CN115798055A (en) | Violent behavior detection method based on corersort tracking algorithm | |
| CN111814581A (en) | Student behavior identification method and system based on classroom scene | |
| Wang et al. | Lightweight Network Based Real-time Anomaly Detection Method for Caregiving at Home | |
| CN110598569A (en) | Action recognition method based on human body posture data | |
| CN116824460B (en) | Face recognition-based examinee track tracking method, system and medium | |
| Dou et al. | Blink Detection Based on Pixel Fluctuation Ratio of Eye Image | |
| Huang et al. | A Pilot Study of Deep Learning Models for Camera based Hand Hygiene Monitoring in ICU |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |