CN114678117A - Management method and device for standardizing operating behaviors of operating room personnel - Google Patents
Management method and device for standardizing operating behaviors of operating room personnel Download PDFInfo
- Publication number
- CN114678117A CN114678117A CN202210581702.8A CN202210581702A CN114678117A CN 114678117 A CN114678117 A CN 114678117A CN 202210581702 A CN202210581702 A CN 202210581702A CN 114678117 A CN114678117 A CN 114678117A
- Authority
- CN
- China
- Prior art keywords
- model
- operating room
- data
- recognition model
- operating
- 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
- 238000007726 management method Methods 0.000 title claims abstract description 59
- 230000006399 behavior Effects 0.000 title claims abstract description 34
- 238000004458 analytical method Methods 0.000 claims abstract description 45
- 238000000034 method Methods 0.000 claims abstract description 39
- 230000008569 process Effects 0.000 claims abstract description 31
- 238000013145 classification model Methods 0.000 claims abstract description 17
- 238000012549 training Methods 0.000 claims abstract description 13
- 238000009826 distribution Methods 0.000 claims abstract description 4
- 238000004422 calculation algorithm Methods 0.000 claims description 30
- 238000001514 detection method Methods 0.000 claims description 19
- 238000005406 washing Methods 0.000 claims description 15
- 206010002091 Anaesthesia Diseases 0.000 claims description 13
- 230000037005 anaesthesia Effects 0.000 claims description 13
- 238000004590 computer program Methods 0.000 claims description 12
- 230000000875 corresponding effect Effects 0.000 claims description 10
- 229940124645 emergency medicine Drugs 0.000 claims description 10
- 238000012545 processing Methods 0.000 claims description 10
- 238000004140 cleaning Methods 0.000 claims description 8
- 239000013598 vector Substances 0.000 claims description 8
- 238000007405 data analysis Methods 0.000 claims description 6
- 239000003814 drug Substances 0.000 claims description 6
- 238000007619 statistical method Methods 0.000 claims description 6
- 230000002596 correlated effect Effects 0.000 claims description 5
- 238000012360 testing method Methods 0.000 claims description 5
- 208000015181 infectious disease Diseases 0.000 claims description 4
- 238000003062 neural network model Methods 0.000 claims description 4
- 244000052769 pathogen Species 0.000 claims description 4
- 230000001717 pathogenic effect Effects 0.000 claims description 4
- 238000012806 monitoring device Methods 0.000 claims description 2
- 238000003860 storage Methods 0.000 claims description 2
- 238000000265 homogenisation Methods 0.000 abstract description 5
- 201000010099 disease Diseases 0.000 abstract description 2
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 abstract description 2
- 238000010276 construction Methods 0.000 description 5
- 238000013527 convolutional neural network Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000002372 labelling Methods 0.000 description 3
- 230000007704 transition Effects 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 230000004927 fusion Effects 0.000 description 2
- 230000014509 gene expression Effects 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 230000007935 neutral effect Effects 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000019771 cognition Effects 0.000 description 1
- 238000005094 computer simulation Methods 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 210000005069 ears Anatomy 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 230000009191 jumping Effects 0.000 description 1
- 238000007477 logistic regression Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 230000002980 postoperative effect Effects 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 238000013441 quality evaluation Methods 0.000 description 1
- 238000013139 quantization Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 230000033772 system development Effects 0.000 description 1
- 230000009897 systematic effect Effects 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
Images
Classifications
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/367—Ontology
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/02—Feature extraction for speech recognition; Selection of recognition unit
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/06—Creation of reference templates; Training of speech recognition systems, e.g. adaptation to the characteristics of the speaker's voice
- G10L15/063—Training
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computational Linguistics (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- General Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Data Mining & Analysis (AREA)
- Business, Economics & Management (AREA)
- General Business, Economics & Management (AREA)
- Acoustics & Sound (AREA)
- General Physics & Mathematics (AREA)
- Databases & Information Systems (AREA)
- Biomedical Technology (AREA)
- Medical Informatics (AREA)
- Primary Health Care (AREA)
- Public Health (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Animal Behavior & Ethology (AREA)
- Epidemiology (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Medical Treatment And Welfare Office Work (AREA)
Abstract
The invention discloses a management method and a device for standardizing operating behaviors of operating room personnel, wherein the method comprises the steps of S1, constructing a voice recognition model, an image target recognition model and a data dynamic analysis model, and constructing an operating room clean content classification model by using a first knowledge graph; s2, training a voice recognition model image target recognition model and a data dynamic analysis model; s3 constructing a second knowledge graph; s4, combining the trained voice recognition model, the trained image target recognition model, the trained data dynamic analysis model, the trained operating room clean content classification model and the trained second knowledge map to form a management system; s5, managing the distribution and operation of the personnel in the operating room; through the standardized operation process of operating medical personnel, the obvious differences of different departments and different patients in the purposes and ranges of disease types, illness state relief, operation modes and control of the same index are eliminated, and the homogenization of medical resources and the homogenization of medical quality are promoted.
Description
Technical Field
The invention relates to the field of artificial intelligence, in particular to a management method and a management device for standardizing operating behaviors of operating room personnel.
Background
In China, the standardization and implementation of operation in an operating room are a common subject in clinical work, and due to the lack of implementation of an important operating room image informatization system in the prior art, most of the construction of the digital operating room in China is mainly based on operation preparation teaching and equipment control. And the following false recognitions still exist in China: some hospitals consider that the digital operating room is equivalent to an operation teaching and transmission system, and some hospitals consider that the anesthesia information system is equipped to be equal to the digital operating room, and even consider that the digital operating room is necessarily expensive to construct and needs a large scale. More importantly, because different hospitals have larger differences in cognition of operation management, standardized operation and the like, the compatibility and the normalization of the system are restricted; therefore, the existing technology not only lacks systematic and standardized management on the operation procedure of the operating room, but also fails to solve the problems faced by operating room operation standardization and homogenization of medical services.
Disclosure of Invention
The invention aims to solve the problems and designs a management method and a management device for standardizing operating behaviors of operating room personnel.
The invention realizes the purpose through the following technical scheme:
a management method for standardizing operating behaviors of operating room personnel comprises the following steps:
s1, constructing a voice recognition model, a plurality of image target recognition models and a plurality of data dynamic analysis models, and constructing an operating room clean content classification model by using a first knowledge graph, wherein the first knowledge graph is the corresponding relation between the pathogen infection condition of a patient and the subsequent clean content of instruments, consumables, an operating room and an operator;
s2, acquiring a test data set related to operating behaviors of the operating room, and respectively training a voice recognition model, an image target recognition model and a data dynamic analysis model by using a voice recognition algorithm, an image target detection algorithm and a data analysis algorithm;
s3, acquiring a standardized management process formulated and recognized at home and abroad at present, and constructing a second knowledge graph by adopting a knowledge graph algorithm;
s4, receiving recognition results of the trained voice recognition model, the trained image target recognition model, the trained data dynamic analysis model and the operating room clean content classification model by the second knowledge map, and combining the first knowledge map, the second knowledge map, the trained voice recognition model, the trained image target recognition model, the trained data dynamic analysis model and the operating room clean content classification model to form a management system;
and S5, acquiring an operation arrangement form, collecting data related to the operation in real time, importing the data into a management system, connecting the data with the operation video content identification model, and managing the distribution and operation of the personnel in the operation room according to the current medical standard by combining the arrangement and the operation progress of different operation rooms.
Management device for standardizing operating actions of operating room personnel, comprising:
an acquisition module; the acquisition module is used for acquiring data related to the operation;
a processor; the processor is used for running a computer program, the data signal output end of the acquisition module is connected with the data signal input end of the processor, and the computer program realizes the steps of the management method for standardizing the operation behaviors of the operating room personnel when being executed by the processor;
a reservoir; the memory is used for storing the data acquired by the acquisition module, the computer program and the violation analysis result obtained by the processor executing the computer program, the data signal output end of the acquisition module is connected with the data signal input end of the memory, and the signal end of the memory is connected with the signal end of the processor.
The invention has the beneficial effects that: through the standardized management of the operation and the flow of the medical staff, the obvious differences of different departments and different patients in the purposes and ranges of disease types, illness states, operation modes and the control of the same index are further eliminated, and the homogenization of medical resources and the homogenization of medical quality are promoted. Meanwhile, through the management of the AI model on the operation process of the medical staff, the subsequently developed model for managing the medical staff in a standardized way can be used universally among different regions and departments, so that the technical development cost of the relevant model is reduced, the compatibility and the standardization of the system are improved, the personnel management efficiency of the operating room is improved, the system development cost is reduced, the standardization of the surgical medical treatment is enhanced, the surgical operation quality is further improved, and the obstacles faced by the standardized digital management of the operating room in China are accelerated.
Drawings
FIG. 1 is a flow chart of the steps of the present invention;
FIG. 2 is a schematic diagram of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, 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. It is to be understood that the embodiments described are only a few embodiments of the present invention, and not all embodiments. The components of embodiments of the present invention generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations.
Thus, the following detailed description of the embodiments of the present invention, as presented in the figures, is not intended to limit the scope of the invention, as claimed, but is merely representative of selected embodiments of the invention. All other embodiments, which can be obtained by a person skilled in the art without inventive step based on the embodiments of the present invention, are within the scope of protection of the present invention.
It should be noted that: like reference numbers and letters refer to like items in the following figures, and thus, once an item is defined in one figure, it need not be further defined and explained in subsequent figures.
In the description of the present invention, it is to be understood that the terms "upper", "lower", "inside", "outside", "left", "right", and the like, indicate orientations or positional relationships based on the orientations or positional relationships shown in the drawings, or the orientations or positional relationships that the products of the present invention are conventionally placed in use, or the orientations or positional relationships that are conventionally understood by those skilled in the art, and are used for convenience of describing the present invention and simplifying the description, but do not indicate or imply that the devices or elements referred to must have a specific orientation, be constructed in a specific orientation, and be operated, and thus, should not be construed as limiting the present invention.
Furthermore, the terms "first," "second," and the like are used merely to distinguish one description from another, and are not to be construed as indicating or implying relative importance.
In the description of the present invention, it is also to be noted that, unless otherwise explicitly stated or limited, the terms "disposed" and "connected" are to be interpreted broadly, and for example, "connected" may be a fixed connection, a detachable connection, or an integral connection; can be mechanically or electrically connected; the connection may be direct or indirect via an intermediate medium, and may be a communication between the two elements. The specific meanings of the above terms in the present invention can be understood by those skilled in the art according to specific situations.
The following detailed description of embodiments of the invention refers to the accompanying drawings.
As shown in fig. 1, the management method for standardizing the operation behavior of the operating room personnel comprises the following steps:
s1, constructing a voice recognition model, a plurality of image target recognition models and a plurality of data dynamic analysis models, constructing an operating room clean content classification model by using a first knowledge graph, accessing medical record information, inspection results and a proposed operation mode of a patient in real time, and prompting medical staff to clean different materials subsequently after the operation is finished in real time, wherein the first knowledge graph is the corresponding relation between the pathogen infection condition of the patient and the subsequent clean content of instruments, consumables, an operation room and an operator;
s2, acquiring a test data set related to operating behaviors of the operating room, and respectively training a voice recognition model, an image target recognition model and a data dynamic analysis model by using a voice recognition algorithm, an image target detection algorithm and a data analysis algorithm; the test data set comprises operating room audio data, video image data, the electrocardio information of the electrocardio monitoring device and anesthesia information of the whole course of an operation, the video image data comprises videos of an operating room door, a changing room access and exit, a hand washing table, an operating room corner, an operating table overhead and an operating room instrument placing table and a medicine cabinet, and the training model specifically comprises:
importing the audio data of the operating room into a voice recognition model, training by adopting a voice recognition algorithm to obtain the voice recognition model in the operating room, constructing a word cloud database related to the work of the operating room, and corresponding the data of the word cloud database with the output result of the voice recognition model in the operating room to form a voice recognition analysis model in the operating room;
importing video image data into a plurality of image target recognition models, and training by adopting an image target detection algorithm to obtain an operating room doorway person recognition counting model, a dressing result recognition model, a hand washing state recognition model, an operating table instrument recognition model, an operating room person behavior recognition model, a doctor position recognition model and an emergency medicine recognition model;
the electrocardiogram information and the anesthesia information are respectively imported into the two data dynamic analysis models, and a data analysis algorithm is adopted to obtain a vital sign dynamic analysis model and an anesthesia information analysis model.
S3, acquiring the current standardized management flow formulated and recognized at home and abroad, and constructing a second knowledge graph by adopting a knowledge graph algorithm.
S4, receiving recognition results of the trained voice recognition model, the trained image target recognition model, the trained data dynamic analysis model and the operating room clean content classification model by the second knowledge map, and combining the first knowledge map, the second knowledge map, the trained voice recognition model, the trained image target recognition model, the trained data dynamic analysis model and the operating room clean content classification model to form a management system; the method specifically comprises the following steps:
the recognition result of the voice recognition analysis model in the operating room, the dynamic vital sign analysis model, the anesthesia information analysis model and the analysis result of the cleaning content classification model in the operating room are correlated with the recognition results of the personnel behavior recognition model in the operating room, the operating table instrument recognition model, the doctor position recognition model and the first-aid medicine recognition model through the computer neural network model in real time, the correlated correlation result, the operating room entrance guard information, the operator recognition counting model at the operating room entrance, the dressing result recognition model and the hand washing state recognition model are output to be used as the input of a second knowledge map to form a management system, and the second knowledge map analyzes to obtain the conformity between the recognized information and the content specified by the management process.
And S5, acquiring an operation arrangement form, collecting data related to the operation in real time, importing the data into a management system, connecting the data with the operation video content identification model, and managing the distribution and operation of the personnel in the operation room according to the current medical standard by combining the arrangement and the operation progress of different operation rooms.
The voice recognition model comprises a signal processing module and a decoder module, wherein the signal processing module is used for extracting voice characteristics and converting the voice characteristics into characteristic vector signals, the decoder module comprises an acoustic model and a language model, the acoustic model adopts a hidden Markov model for modeling, the language model adopts an N-element grammar for modeling, the output of the signal processing module is used as the input of the acoustic model, and the output of the acoustic model is used as the input of the language model.
The language identification model is composed of a signal processing module and a decoder module. The signal processing module extracts the most important features in the voice according to the auditory perception characteristics of human ears and converts the voice signals into feature vector sequences. The decoder converts the input speech feature vector sequence into a character sequence according to the acoustic model and the language model. Wherein the decoder module is composed of an acoustic model and a language model. Acoustic models are knowledge representations of the acoustics, phonetics, variables of the environment, as well as differences in speaker gender, accents, and the like. A language model is a knowledge representation of a set of word sequences.
The acoustic model in the speech recognition makes full use of information such as acoustics, phonetics, environmental characteristics, speaker-specific accents and the like to model the speech. Current speech recognition systems often use Hidden Markov Model (HMM) modeling to represent the a posteriori probability of a sequence of speech feature vectors to a sequence of states. Hidden Markov models are probabilistic graphical models that can be used to represent correlations between sequences, often used to model temporal data.
A hidden markov model is a weighted directed graph, with each node on the graph called a state. At each moment, the hidden Markov model jumps from one state to another with a certain probability, and transmits an observation symbol with a certain probability, and the probability of jumping is represented by the weight on the edge. The hidden Markov model assumption is that each state transition is only related to the previous state and is not related to other states before and after, namely the Markov assumption; the symbols transmitted in each state are only relevant to the current state and have no relation to other states and other symbols, i.e. hypotheses are output independently. A hidden markov model is generally represented by a triplet γ = (a, B, pi), where a is a state transition probability matrix representing the probability of transitioning to another state at a state; b is a symbol probability matrix which represents the probability of transmitting a certain symbol in a certain state; and pi is an initial state probability vector and represents the probability of being in a certain state at the beginning. The hidden Markov model can generate two random sequences, one is a state sequence and the other is an observation symbol sequence, so that the hidden Markov model is a double random process, but the outside can only observe the observation symbol sequence and cannot observe the state sequence. The Viterbi Algorithm (Viterbi Algorithm) can be used to find the state sequence that occurs with the highest probability given the observed symbol sequence. The probability of a certain observed symbol sequence can be efficiently obtained by a Forward-Backward Algorithm (Forward-Backward Algorithm). The transition probability and observed symbol emission probability for each state can be calculated by the Baum-Welch Algorithm (Baum-Welch Algorithm).
Hidden markov models are typically used in speech recognition to model the relationship between acoustic elements and speech feature sequences. Generally, the level of an acoustic unit is small, the number is small, but the sensitivity to context is large. Sub-words (Sub-words) are generally used as acoustic units in a large-vocabulary continuous speech recognition system, for example, phonemes are used in english, initials and finals are used in chinese, and the like. The topological structure of the hidden Markov model in the acoustic model generally adopts a three-state structure from left to right, and each state has an arc pointing to the state. Since continuous speech has a phenomenon of co-articulation, the three preceding and following phonemes need to be considered together, which is called a Triphone (Triphone) model. After the triphone is introduced, the number of hidden Markov models is increased sharply, so that the states are generally clustered, and the clustered states are called Senone. The acoustic feature vector values in the speech recognition task are continuous, and in order to eliminate errors caused by the quantization process, the probability of the state of the feature vector is considered to be modeled by using a continuous probability density function.
A language model may represent the probability of a certain sequence of words occurring. A common language model in speech recognition is N-Gram (N-Gram), which is the probability of the occurrence of N words before and after statistics. N-gram assumes that the probability of a word occurring is only related to the probability of the previous N-1 words occurring.
Let us now have a word sequence W = (W)1, w2, ww=(w1, w2,…, wu) W represents a complete word sequence, W1, w2,…, wuRepresenting a single word sequence, its probability of occurrence can be decomposed into the following form:
wherein P (W) represents the probability of occurrence of a complete word sequence, P (W)n) Indicates the probability of occurrence of the nth word sequence, P (W)n| Wn-1) Representing the probability that the nth word sequence occurs under the conditions of the (n-1) th word sequence.
However, such probabilities cannot be counted. According to the Markov assumption, only the probability under the condition of generating the first N characters is considered. Suppose that N =2 has
Then according to the Bayes formula, the probability of a certain word occurring under the condition of another word can be obtained
Therefore, the probability of occurrence of adjacent words is counted in a large amount of linguistic data, and then the probability of occurrence of a single word is counted. Since some rare phrases are not present in the corpus but have a probability of occurrence, the probability of generating the rare phrases by the algorithm, i.e. smoothing, is required.
The image target identification model is formed by sequentially connecting a backhaul, a neutral and a Head, wherein the backhaul is used for extracting the characteristic information of video image data, the neutral is used for fusing the characteristic information extracted by the backhaul, and the Head is used for predicting the position and the type of a target by using the characteristic information extracted by the backhaul.
Backbone network. The backhaul serves as part of the overall target detection network. The backhaul refers to a feature extraction network in the whole target detection network, and the function of the backhaul is to extract feature information in a picture. For example, a Convolutional Neural Network (CNN) is used to extract features of an inputted picture, common points between the features are extracted, and the feature map size is reduced by continuous convolution, thereby finding the most core portion.
The Neck part is the key link for starting and stopping in the target detection framework. The method has the advantages that the Neck mainly integrates the characteristics extracted by the backhaul in the target detection network, so that the characteristics learned by the network are more diverse, and the performance of the detection network is improved. The characteristics given by the backhaul are better fused and extracted, and then the characteristics are transmitted to the subsequent Head for detection, so that the network performance is improved. And the Neck processes and reasonably utilizes the important characteristics extracted by the Backbone, thereby being beneficial to the next step of learning a specific task of Head. The heck is placed between the backbone and the head in order to better utilize the features extracted by the backbone.
Head, which is generally called a detection header in the object detection network. Head is the network that takes the network output content and uses the previously extracted features that the Head uses to make predictions. Head may be understood as the feature extracted from Backbone from which the location and class of the target is predicted. In addition to identifying the category of the object, the target detection also needs to locate the object, and the main function is to locate and classify the object.
According to the difference of the position and the content of the collected video picture data and the difference of application scenes, the models are also divided into an operating room door person identification counting model, a dressing result identification model, a hand washing state identification model, an operating table instrument identification model, an operating room person behavior identification model, a doctor position identification model, an emergency medicine identification model and the like. The operating table instrument recognition model, the doctor position recognition model, the first-aid medicine recognition model and the like also perform classified counting on the recognized contents.
Data dynamic analysis models include, but are not limited to: the model structure can dynamically analyze multiple groups of scalar data and obtain various different classification conclusions simultaneously, such as a logistic regression model, a ridge regression model, a lasso regression model, an elastic regression model, a decision tree, an artificial neural network, a convolutional neural network and the like.
Training of an image target recognition model, comprising:
selecting pictures containing the required targets from different operation scene pictures, and labeling the position information of each target to obtain a target detection labeling database; the target position information is a minimum circumscribed rectangular frame containing a target in the picture;
training a target detection network according to the target detection labeling database;
after training is finished, the required target detection models, namely an operation room doorway person identification counting model, a dressing result identification model, a hand washing state identification model, an operation table instrument identification model, an operation room person behavior identification model, a doctor position identification model and an emergency medicine identification model are obtained.
The knowledge graph mainly has two construction modes of top-down (top-down) and bottom-up (bottom-up). Top-down refers to defining the ontology and data schema for the knowledge graph and then adding the entity to the knowledge base. The construction method needs to utilize some existing structured knowledge base as a basic knowledge base. And the bottom-up method comprises the steps of extracting entities from some open link data, selecting the entities with higher confidence degrees, adding the entities into a knowledge base, and then constructing a top-level ontology mode.
The process of constructing the knowledge graph in a bottom-up manner is an iterative updating process, and each round of updating comprises 3 steps: 1) information extraction, namely extracting entities (concepts), attributes and interrelations among the entities from various types of data sources, and forming ontology knowledge expression on the basis; 2) knowledge fusion, in which after new knowledge is obtained, it is necessary to integrate the new knowledge to eliminate contradictions and ambiguities, for example, some entities may have multiple expressions, and a specific name may correspond to multiple different entities; 3) and (4) knowledge processing, namely, for the fused new knowledge, after quality evaluation (part of the new knowledge needs to be manually screened), adding the qualified part into a knowledge base to ensure the quality of the knowledge base. After the data is newly added, knowledge reasoning can be carried out, the existing knowledge can be expanded, and new knowledge can be obtained. Specifically, in some embodiments, the construction of the knowledge-graph comprises the following steps:
1. establishing a basic framework of a required knowledge graph according to a universal data standard;
2. uniformly standardizing the relationship among all entities in the basic architecture to obtain a standard dictionary table with standard specifications;
3. obtaining semi-structured data related to content in the knowledge-graph;
4. extracting entity information of key entities from the semi-structured data;
5. performing data fusion on the entity information according to the standard dictionary table to form structured data;
6. and generating corresponding data structure pairs by using the structured data, and storing the data structure pairs as the knowledge graph.
Management device for standardizing operating actions of operating room personnel, comprising:
an acquisition module; the acquisition module is used for acquiring data related to the operation;
a processor; the processor is used for running a computer program, the data signal output end of the acquisition module is connected with the data signal input end of the processor, and the computer program realizes the steps of the management method for standardizing the operation behaviors of the operating room personnel when being executed by the processor;
a reservoir; the memory is used for storing the data acquired by the acquisition module, the computer program and the violation analysis result obtained by the processor executing the computer program, the data signal output end of the acquisition module is connected with the data signal input end of the memory, and the signal end of the memory is connected with the signal end of the processor.
The management device further comprises a statistical analysis module, the statistical analysis module is used for carrying out classification and summary statistics on the violation behaviors in the set time range based on the common knowledge acknowledged by statistics, and a data signal input end of the statistical analysis module is connected with a data signal output end of the storage.
The specific process comprises the following steps:
1. operating room data acquisition
As shown in fig. 2, the doorway of the operating room, the entrance and exit of the changing room, the hand washing table, the corner of the operating room, the upper space of the operating table, the instrument placing table of the operating room and the medicine cabinet are all provided with camera lenses for respectively collecting video pictures of corresponding areas. Meanwhile, the wall corner of the operating room and the upper part of the operating table are also provided with audio acquisition devices to collect audio information of corresponding areas. The audio and video contents of different scenes collected by the audio and video image collecting devices and the access control information of the operating room are sent to the corresponding computer models.
2. Combining and correlating model data
As shown in fig. 2, for video information, the doorway person identification and counting model of the operating room, the dressing result identification model, the hand washing state identification model, the operating table instrument identification model, the personnel behavior identification model of the operating room, the doctor position identification model, the emergency medicine identification model, etc., video information collected by the camera lenses placed at the doorway of the operating room, the doorway of the dressing room, the hand washing table, the surgical instrument placement table, the corner of the operating room, and the medicine cabinet is received, and the camera lenses above the operating table also transmit the video information to the doctor position identification model and the operating table instrument identification model. The operation room doorway person identification model, the dressing result identification model, the hand washing state identification model, the operation table instrument identification model, the operation room personnel behavior identification model, the doctor position identification model and the emergency medicine identification model respectively identify and count the flow of the personnel at the doorway of the operation room, the dressing state of the medical personnel participating in the operation, the hand washing state of the medical personnel participating in the operation, the types and the respective counts of the instruments used in the operation, the behavior of the operation room personnel, the positions and the respective counts of the operation room personnel and the types and the respective counts of the emergency medicines through a trained computer neural network in real time.
For the audio information, the audio collecting devices at the wall corner of the operating room and above the operating table transmit the audio information to the voice recognition analysis model in the operating room in real time.
The vital sign dynamic analysis model, the anesthesia information analysis model and the operating room clean content classification model collect data of the electrocardiographic monitoring, the anesthesia information and the patient medical record information in real time through data lines, broadband networks, radio communication and other information transmission technologies.
The recognition result of the voice content, the dynamic vital sign analysis model, the anesthesia information analysis model and the analysis result of the operating room cleaning content classification model formed by the first knowledge graph are correlated with the recognition results of the operating room personnel behavior model, the operating table instrument recognition model, the doctor position recognition model and the emergency medicine recognition model in real time through the computer neural network model.
3. Construction of first and second knowledge maps and association between first and second knowledge maps and model
Establishing a corresponding relation between basic information such as pathogen infection conditions of patients and subsequent cleaning contents of instruments, consumables, an operating room and operators by using a knowledge graph algorithm and combining current standardized management processes formulated and recognized at home and abroad, including but not limited to contents such as an operating room cleaning process, an operating room instrument consumable processing process and the like, and establishing a first knowledge graph for an operating room cleaning content classification model;
the method comprises the steps of using a knowledge graph algorithm, combining with current standardized management processes formulated and recognized at home and abroad, including but not limited to an operating room personnel flow management process, a hand washing operation standardized management process, a patient preoperative checking standardized management process, an instrument checking standardized management process, an intraoperative emergency management process, an intraoperative rescue standardized management process, an intraoperative instrument desertion standardized management process, an intraoperative personnel behavior standardized management process, a patient postoperative checking standardized management process, an instrument final checking standardized management process, an operating room cleaning standardized management process and the like, and constructing a second knowledge graph for the different models.
The second knowledge graph receives the identification results of 1) the entrance guard information of the operating room, 2) the identification counting model of the personnel at the entrance guard of the operating room, the identification model of the dressing result and the identification result of the hand washing state identification model in real time, and 3) the identification results of the voice content and the dynamic analysis model of the vital signs, the anesthesia information analysis model and the classification model of the cleaning content of the operating room are related to the identification results of the personnel behavior model of the operating room, the instrument identification model of the operating table, the position identification model of the doctor and the identification model of the emergency medicine in real time through the computer neural network model, and the conformity between the information identified by the second knowledge graph and the content specified by the management flow is analyzed. If the identified information does not conform to the regulations for managing the process contents, the model prompts the correct operation mode in real time through the display screens broadcasted and/or configured at the doorway of the operating room, the dressing room, the operating table and the operating room according to the scene violating the regulations, and transmits and stores the contents violating the regulations to a database preset in the computer of an operating room manager or other given paths through a network or a data line.
4. Recording, statistics and analysis of violations
According to the setting of an operating room administrator, such as monthly statistics, quarterly statistics, annual statistics and the like, statistical analysis software or plug-ins built in a model automatically acquire records of rule violations in different operating areas/operating rooms recorded in a database, wherein the records comprise time, places, operating teams and contents of the rule violations, and different types of violation behaviors are classified and summarized based on a statistically accepted method, so that the normative and the change of operating room operation of hospitals and different operating teams in different air are obtained. The operating room manager can also obtain the recommendation of the operating room management optimization through the statistical result.
The technical solution of the present invention is not limited to the limitations of the above specific embodiments, and all technical modifications made according to the technical solution of the present invention fall within the protection scope of the present invention.
Claims (8)
1. Management method for standardizing the operating behaviour of operating room personnel, characterized in that it comprises:
s1, constructing a voice recognition model, a plurality of image target recognition models and a plurality of data dynamic analysis models, and constructing an operating room clean content classification model by using a first knowledge graph, wherein the first knowledge graph is the corresponding relation between the pathogen infection condition of a patient and the subsequent clean content of instruments, consumables, an operating room and an operator;
s2, acquiring a test data set related to operating behaviors of the operating room, and respectively training a voice recognition model, an image target recognition model and a data dynamic analysis model by using a voice recognition algorithm, an image target detection algorithm and a data analysis algorithm;
s3, acquiring a standardized management process, wherein the standardized management process comprises but is not limited to an operating room cleaning process and operating room instrument consumable processing process content, and constructing a second knowledge graph by adopting a knowledge graph algorithm;
s4, receiving recognition results of the trained voice recognition model, the trained image target recognition model, the trained data dynamic analysis model and the operating room clean content classification model by the second knowledge map, and combining the first knowledge map, the second knowledge map, the trained voice recognition model, the trained image target recognition model, the trained data dynamic analysis model and the operating room clean content classification model to form a management system;
and S5, acquiring an operation arrangement form, collecting data related to the operation in real time, importing the data into a management system, connecting the data with the operation video content identification model, and managing the distribution and operation of the personnel in the operation room according to the current medical standard by combining the arrangement and the operation progress of different operation rooms.
2. The method of claim 1, wherein the test data set comprises audio data of the operating room, video image data, electrocardiographic information of the electrocardiographic monitoring device and anesthesia information of the operating room during the whole operation, and the video image data comprises videos of the doorway of the operating room, the entrance and exit of the dressing room, the washing table, the corner of the operating room, the upper space of the operating table, the placing table of the operating room instruments and the medicine cabinet in S2.
3. The management method for standardizing the operation behavior of the operating room personnel as claimed in claim 2, wherein the training of the speech recognition model, the image object recognition model and the data dynamic analysis model using the speech recognition algorithm, the image object detection algorithm and the data analysis algorithm in S2 is specifically:
importing the audio data of the operating room into a voice recognition model, training by adopting a voice recognition algorithm to obtain the voice recognition model in the operating room, constructing a word cloud database related to the work of the operating room, and corresponding the data of the word cloud database with the output result of the voice recognition model in the operating room to form a voice recognition analysis model in the operating room;
importing video image data into a plurality of image target recognition models, and training by adopting an image target detection algorithm to obtain an operating room doorway person recognition counting model, a dressing result recognition model, a hand washing state recognition model, an operating table instrument recognition model, an operating room person behavior recognition model, a doctor position recognition model and an emergency medicine recognition model;
the electrocardio information and the anesthesia information respectively construct two data dynamic analysis models, and a data analysis algorithm is adopted to obtain the results of the vital sign dynamic analysis model and the anesthesia information analysis model.
4. The management method for standardizing the operation behavior of the operating room personnel as claimed in claim 3, wherein in S4, the recognition results of the voice recognition analysis model in the operating room, the dynamic vital sign analysis model, the anesthesia information analysis model and the classification model of the clean content in the operating room are correlated with the recognition results of the personnel behavior recognition model in the operating room, the operating table instrument recognition model, the doctor position recognition model and the emergency medicine recognition model through the computer neural network model in real time, the correlated result, the operating room entrance guard information, the identification counting model of the personnel at the operating room, the output of the dressing result recognition model and the hand washing state recognition model are used as the input of the knowledge map to form the management system, and the knowledge map analysis obtains the conformity degree between the recognized information and the content specified by the management process.
5. The management method for normalizing operating behavior of operating room personnel of claim 3, wherein the speech recognition model comprises a signal processing module for extracting speech features and converting the speech features into feature vector signals and a decoder module comprising an acoustic model and a language model, the acoustic model is modeled using a hidden Markov model, the language model is modeled using an N-gram, an output of the signal processing module is an input of the acoustic model, and an output of the acoustic model is an input of the language model.
6. The management method for standardizing the operation behavior of the personnel in the operating room as recited in claim 3, wherein the image object recognition model is formed by sequentially connecting a Backbone, a Neok and a Head, the Backbone is used for extracting the characteristic information of the video image data, the Neok is used for fusing the characteristic information extracted by the Backbone, and the Head is used for predicting the position and the category of the object by using the fused characteristic information extracted by the Backbone.
7. Management device for standardizing operating actions of operating room personnel, comprising:
an acquisition module; the acquisition module is used for acquiring data related to the operation;
a processor; the processor is used for running a computer program, a data signal output end of the acquisition module is connected with a data signal input end of the processor, and the computer program realizes the steps of the management method for standardizing the operation behaviors of the operating room personnel according to any one of claims 1-6 when being executed by the processor;
a reservoir; the memory is used for storing the data acquired by the acquisition module, the computer program and the violation analysis result obtained by the processor executing the computer program, the data signal output end of the acquisition module is connected with the data signal input end of the memory, and the signal end of the memory is connected with the signal end of the processor.
8. The management device for standardizing the operating behaviors of the personnel in the operating room as claimed in claim 7, wherein the management device further comprises a statistical analysis module, the statistical analysis module is used for carrying out classified summary statistics on the illegal behaviors in the set time range, and a data signal input end of the statistical analysis module is connected with a data signal output end of the storage.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210581702.8A CN114678117B (en) | 2022-05-26 | 2022-05-26 | Management method and device for standardizing operating behaviors of operating room personnel |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210581702.8A CN114678117B (en) | 2022-05-26 | 2022-05-26 | Management method and device for standardizing operating behaviors of operating room personnel |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114678117A true CN114678117A (en) | 2022-06-28 |
| CN114678117B CN114678117B (en) | 2022-08-02 |
Family
ID=82081079
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210581702.8A Active CN114678117B (en) | 2022-05-26 | 2022-05-26 | Management method and device for standardizing operating behaviors of operating room personnel |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114678117B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116091963A (en) * | 2022-12-22 | 2023-05-09 | 广州奥咨达医疗器械技术股份有限公司 | Quality evaluation method and device for clinical test institution, electronic equipment and storage medium |
| CN116312092A (en) * | 2023-03-15 | 2023-06-23 | 广东爱唯视医疗科技有限公司 | Teaching method of ophthalmic surgery teaching platform based on 5G-VR |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2001095253A2 (en) * | 2000-06-09 | 2001-12-13 | Medimage Aps | Quality and safety assurance of medical applications |
| CN106373062A (en) * | 2016-08-31 | 2017-02-01 | 浙江金华广福肿瘤医院 | Operation process management system and perioperative period standardization work process |
| CN106528277A (en) * | 2015-09-14 | 2017-03-22 | 北京空间飞行器总体设计部 | Standardized scheduling method applicable to autonomous FDIR of large SAR load |
| CN109166606A (en) * | 2018-08-31 | 2019-01-08 | 上海轩昂医疗科技有限公司 | A kind of electronic health record online editing platform and its implementation |
| CN111667902A (en) * | 2019-03-08 | 2020-09-15 | 震一科技(天津)有限公司 | Clinical operation management system |
| CN112415921A (en) * | 2020-10-12 | 2021-02-26 | 深圳市豫知科技有限公司 | Iatrogenic infection and medical adverse event tracing system based on whole-process monitoring |
| CN112506123A (en) * | 2020-10-12 | 2021-03-16 | 深圳市豫知科技有限公司 | Medical diagnosis and treatment environment monitoring system aiming at clean requirements |
| WO2021097367A1 (en) * | 2019-11-15 | 2021-05-20 | Onsite-Llc | Monitoring and enforcing infection safety procedures in operating rooms |
| CN112932663A (en) * | 2021-03-02 | 2021-06-11 | 成都与睿创新科技有限公司 | Intelligent auxiliary method and system for improving safety of laparoscopic cholecystectomy |
| CN112966903A (en) * | 2021-02-04 | 2021-06-15 | 成都鼎安华智慧物联网股份有限公司 | Dangerous chemical safety production risk monitoring and early warning system and method |
| WO2021185049A1 (en) * | 2020-03-14 | 2021-09-23 | 厦门波耐模型设计有限责任公司 | Medical service robot device, and method and system therof |
| CN114121239A (en) * | 2021-10-21 | 2022-03-01 | 白杨智慧医疗信息科技(北京)有限公司 | Operating room abnormal condition processing method and device and electronic equipment |
-
2022
- 2022-05-26 CN CN202210581702.8A patent/CN114678117B/en active Active
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2001095253A2 (en) * | 2000-06-09 | 2001-12-13 | Medimage Aps | Quality and safety assurance of medical applications |
| CN106528277A (en) * | 2015-09-14 | 2017-03-22 | 北京空间飞行器总体设计部 | Standardized scheduling method applicable to autonomous FDIR of large SAR load |
| CN106373062A (en) * | 2016-08-31 | 2017-02-01 | 浙江金华广福肿瘤医院 | Operation process management system and perioperative period standardization work process |
| CN109166606A (en) * | 2018-08-31 | 2019-01-08 | 上海轩昂医疗科技有限公司 | A kind of electronic health record online editing platform and its implementation |
| CN111667902A (en) * | 2019-03-08 | 2020-09-15 | 震一科技(天津)有限公司 | Clinical operation management system |
| WO2021097367A1 (en) * | 2019-11-15 | 2021-05-20 | Onsite-Llc | Monitoring and enforcing infection safety procedures in operating rooms |
| WO2021185049A1 (en) * | 2020-03-14 | 2021-09-23 | 厦门波耐模型设计有限责任公司 | Medical service robot device, and method and system therof |
| CN112415921A (en) * | 2020-10-12 | 2021-02-26 | 深圳市豫知科技有限公司 | Iatrogenic infection and medical adverse event tracing system based on whole-process monitoring |
| CN112506123A (en) * | 2020-10-12 | 2021-03-16 | 深圳市豫知科技有限公司 | Medical diagnosis and treatment environment monitoring system aiming at clean requirements |
| CN112966903A (en) * | 2021-02-04 | 2021-06-15 | 成都鼎安华智慧物联网股份有限公司 | Dangerous chemical safety production risk monitoring and early warning system and method |
| CN112932663A (en) * | 2021-03-02 | 2021-06-11 | 成都与睿创新科技有限公司 | Intelligent auxiliary method and system for improving safety of laparoscopic cholecystectomy |
| CN114121239A (en) * | 2021-10-21 | 2022-03-01 | 白杨智慧医疗信息科技(北京)有限公司 | Operating room abnormal condition processing method and device and electronic equipment |
Non-Patent Citations (6)
| Title |
|---|
| HARISH RAVI PRAKASH等: "Automatic response assessment in regions of language cortex in epilepsy patients using ECoG-based functional mapping and machine learning", 《2017 IEEE INTERNATIONAL CONFERENCE ON SYSTEMS, MAN, AND CYBERNETICS (SMC)》 * |
| 祁宝莲 等: "基于卷积神经网络的半监督手术视频流程识别", 《计算机科学》 * |
| 赵婧伶 等: "外科洗手及手消毒规范性操作调查分析与对策", 《华西医学》 * |
| 赵霞: "医疗卫生信息标准开发方法学研究与应用", 《中国博士学位论文全文数据库 医药卫生科技辑》 * |
| 郝昕: "基于LSTM的车间生产异常预测与处理方法研究", 《中国优秀博硕士学位论文全文数据库(硕士) 信息科技辑》 * |
| 马文杰 等: "基于依存句法分析与五防操作规范的变电运行操作知识图谱构建", 《科技风》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116091963A (en) * | 2022-12-22 | 2023-05-09 | 广州奥咨达医疗器械技术股份有限公司 | Quality evaluation method and device for clinical test institution, electronic equipment and storage medium |
| CN116091963B (en) * | 2022-12-22 | 2024-05-17 | 广州奥咨达医疗器械技术股份有限公司 | Quality evaluation method and device for clinical test institution, electronic equipment and storage medium |
| CN116312092A (en) * | 2023-03-15 | 2023-06-23 | 广东爱唯视医疗科技有限公司 | Teaching method of ophthalmic surgery teaching platform based on 5G-VR |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114678117B (en) | 2022-08-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114678117B (en) | Management method and device for standardizing operating behaviors of operating room personnel | |
| WO2021233112A1 (en) | Multimodal machine learning-based translation method, device, equipment, and storage medium | |
| US20210034813A1 (en) | Neural network model with evidence extraction | |
| EP3567605A1 (en) | Structured report data from a medical text report | |
| US20200265953A1 (en) | Identifying Valid Medical Data for Facilitating Accurate Medical Diagnosis | |
| CN110337645A (en) | The processing component that can be adapted to | |
| US11468989B2 (en) | Machine-aided dialog system and medical condition inquiry apparatus and method | |
| Wazalwar et al. | Interpretation of sign language into English using NLP techniques | |
| WO2021208444A1 (en) | Method and apparatus for automatically generating electronic cases, a device, and a storage medium | |
| CN113593709B (en) | Disease coding method, system, readable storage medium and device | |
| CN112115240B (en) | Classification processing method, device, server and storage medium | |
| CN109741824B (en) | Medical inquiry method based on machine learning | |
| CN113314100B (en) | Method, device, equipment and storage medium for evaluating and displaying results of spoken language test | |
| CN111986793B (en) | Diagnosis guiding processing method and device based on artificial intelligence, computer equipment and medium | |
| US20240363255A1 (en) | Emergency treatment system, emergency treatment method, and electronic device | |
| US8423498B2 (en) | System and associated method for determining and applying sociocultural characteristics | |
| CN115221941A (en) | Cognitive disorder detection method and related device, electronic equipment and storage medium | |
| CN117390145A (en) | Automatic text dialogue method, device, equipment and medium for clinical test | |
| WO2022015005A1 (en) | Conversation-based mental disorder screening method and device | |
| CN113658690A (en) | Intelligent medical guide method and device, storage medium and electronic equipment | |
| CN116994695A (en) | Training method, device, equipment and storage medium of report generation model | |
| CN116842143A (en) | Dialog simulation method and device based on artificial intelligence, electronic equipment and medium | |
| CN116258136A (en) | Error detection model training method, medical image report detection method, system and equipment | |
| US11475222B2 (en) | Automatically extending a domain taxonomy to the level of granularity present in glossaries in documents | |
| CN113761899A (en) | Medical text generation method, device, equipment and storage medium |
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 |