A protective skill integrated practical training examination system

Abstract

The present application relates to the technical field of medical protection, and more particularly to a protective skill integrated practical training examination system. The present application comprises an AI video acquisition module, a fluorescent pretreatment and imaging module, a fluorescent pollution automatic analysis module, a time sequence correlation module and a scoring module. The present application aims to provide a protective skill integrated practical training examination system to solve the problems of high difficulty and low detection accuracy of existing examination and identification.

Description

Technical Field

[0001] This invention relates to the field of medical protective technology, and in particular to an integrated training and assessment system for protective skills. Background Technology

[0002] In recent years, the expansion of human activity, changes in living environments, and rapid mutation of microorganisms have led to frequent outbreaks of new infectious diseases and their accelerated spread, posing a serious threat to global public health security. my country is one of the countries most severely affected by infectious disease outbreaks. Infectious disease prevention and control, as well as treatment, are important peacetime operations for military medical departments. Under the new circumstances, higher demands are placed on the protective skills of new-type health support personnel.

[0003] Personal protective equipment (PPE) is the last line of defense protecting healthcare workers from exposure to pathogens, and proper donning and doffing is crucial for reducing the risk of occupational infection. However, domestic and international experience shows that the PPE use qualification rate among healthcare workers is generally low, and improper handling during the doffing process leading to self-contamination is a major cause of infection among medical personnel. Current PPE skills training and assessment systems have the following main deficiencies: 1. Traditional assessments rely on instructors' on-site observation and scoring, which suffers from problems such as inconsistent scoring standards, large errors, low efficiency, inability to quantify hidden contamination, and difficulty in fully capturing instantaneous violations.

[0004] 2. In the existing intelligent assessment system, AI action recognition and fluorescent pollution detection operate independently, which cannot achieve correlation analysis between violations and pollution results. It cannot accurately locate the polluting actions, nor can it eliminate environmental interference, resulting in insufficient accuracy of assessment results and a lack of targeted error correction guidance.

[0005] 3. Existing systems generally use fixed scoring weights, which cannot be adaptively adjusted according to the differences in risk of different protection levels and the differences in skill levels of different trainees, resulting in insufficient rationality of the assessment.

[0006] Therefore, designing an integrated training and assessment system for protective skills that can solve the above-mentioned technical problems is a technical issue that needs to be addressed. Summary of the Invention

[0007] To address the aforementioned problems, the present invention aims to provide an integrated training and assessment system for protective skills, thereby resolving the issues of high assessment difficulty and low accuracy in assessment and detection in existing systems.

[0008] To achieve the above objectives, the present invention adopts the following technical solution: including an AI video acquisition module, a fluorescence preprocessing and imaging module, an automatic fluorescence contamination analysis module, a time-series correlation module, and a scoring module; The AI ​​video acquisition module is used to simultaneously acquire PPE donning and doffing action videos, divide human body movements into three regions—upper limbs, torso, and lower limbs—for regional recognition, add timestamps to each action and violation record, and calculate the standardization score of each region's actions. The fluorescence pretreatment and imaging module is used to prepare fluorescent tracers and spray them onto the outer surface of protective clothing, perform ultraviolet imaging on personnel after removal, and transmit the generated human ultraviolet image to the automatic fluorescence pollution analysis module. The automatic fluorescence pollution analysis module is used to receive ultraviolet images of the human body, perform pollution analysis on the upper and lower body of the human body in different regions, and obtain the number of pollution points and pollution area in each region. The time-series correlation module is connected to the AI ​​video acquisition module and the fluorescent pollution automatic analysis module, respectively. It is used to match the timestamp of each violation action collected by the AI ​​video acquisition module with the pollution area, determine the effective pollution caused by the action and eliminate the false pollution remaining in the environment. The scoring module is used to determine the trainee's level based on the assessment's protection registration and historical assessment records, and to calculate the total assessment score.

[0009] Furthermore, the AI ​​video acquisition module includes at least three AI cameras and an edge computing box. It uses a key point human pose model and a key point hand model for action recognition, and uses a confidence-weighted Kalman filter fusion algorithm to output global human key point coordinates.

[0010] Furthermore, the AI ​​video acquisition module uses a bidirectional LSTM network to analyze key point sequences and action temporal features, distinguishing between normal and illegal actions.

[0011] Furthermore, the automatic fluorescence contamination analysis module specifically includes the following steps: First, Gaussian filtering for noise reduction and histogram equalization are used for image preprocessing. Secondly, the nested U-Net algorithm is used to segment the ultraviolet image into upper body and lower body regions; Secondly, a connected component algorithm is used to identify fluorescent residual regions and remove isolated noise points with an area smaller than a preset value. Finally, a kernel density estimation algorithm was used to generate a heat map of human body pollution.

[0012] Furthermore, the time-series correlation module uses the HMM model to automatically divide the PPE removal process into eight time-series segments and establishes a mapping relationship between the time-series segments and the corresponding high-contamination-risk sub-areas within the time-series segments; wherein the time-series segments include removing goggles, removing outer gloves, removing protective clothing, removing inner gloves, removing shoe covers, removing hats, removing masks, and hand hygiene.

[0013] Furthermore, the timing correlation module determines the effective pollution caused by the action and eliminates the false pollution remaining in the environment, specifically including the following steps: Step S1: Set a causal relationship time window centered on the occurrence time of each violation record; Step S2: Calculate the Euclidean distance between the left side of the contamination point to be determined and the operating area corresponding to the violation action; Step S3: Combine the temporal matching degree and spatial matching degree to calculate the confidence level of the causal association between the contamination point and the violation. Step S4: If the confidence level of causal association is greater than the preset value, the contamination point is determined to be a valid contamination caused by the action and is included in the deduction score; if the confidence level of causal association is less than the preset value, it is automatically removed.

[0014] Furthermore, the specific calculation formula for the scoring module is as follows: Final total score = Action standardization score × W1 + Pollution control score × W2 Wherein, W1+W2=1; the basic weights are set according to the protection level: Level 1 protection W1=0.6, W2=0.4, Level 2 protection W1=0.5, W2=0.5, Level 3 protection W1=0.4, W2=0.6.

[0015] Furthermore, the scoring module uses a clustering algorithm to automatically classify trainees into different levels based on historical assessment data, and adjusts the coefficients by adding weights to the base weights for each level. The present invention has the following beneficial effects: 1-This invention divides human movements into three regions: upper limbs, trunk, and lower limbs, and divides contamination into two corresponding regions: upper body and lower body. Through regional movement recognition and regional contamination analysis, it can accurately locate which region's violation caused the contamination in the corresponding region, providing a basis for targeted error correction.

[0016] 2. This invention adds a timestamp to each violation, sets a causal relationship time window, and calculates the causal confidence of pollution by combining time matching degree and spatial matching degree. It can accurately determine the effective pollution caused by the action and automatically eliminate false pollution caused by environmental residues, thereby improving the accuracy and credibility of the assessment results.

[0017] 3- The present invention automatically classifies trainees into levels based on protection level and trainees’ historical assessment records, and adaptively adjusts the scoring weights of action standardization and pollution control to make the assessment standards more in line with the risk requirements of different scenarios and the skill levels of different personnel. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the present invention. Detailed Implementation

[0019] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments: See Figure 1 As shown, the solution includes an AI video acquisition module, a fluorescence preprocessing and imaging module, an automatic fluorescence pollution analysis module, a time-series correlation module, and a scoring module. The AI ​​video acquisition module is used to simultaneously acquire PPE donning and doffing videos, dividing human movements into three regions: upper limbs, torso, and lower limbs for regional recognition. It adds a timestamp to each movement and violation record and calculates a standardization score for each region. Specifically, the base score for standardization is 60 points, with 20 points each for the upper limbs, torso, and lower limbs. Each region initially has 20 points, with points deducted in tiers based on the severity of the violation (2 points for minor violations, 3 points for moderate violations, and 4 points for severe violations). Missing steps or skipping steps deduct 5 points per step, directly from the corresponding region's score. Exceeding the time limit (total duration greater than 15 minutes) deducts 5 points, evenly distributed across the three regions.

[0020] The fluorescence pretreatment and imaging module is used to prepare fluorescent tracers and spray them onto the outer surface of protective clothing, perform ultraviolet imaging on personnel after removal, and transmit the generated human ultraviolet image to the automatic fluorescence pollution analysis module. The automatic fluorescence pollution analysis module is used to receive ultraviolet images of the human body, perform pollution analysis on the upper and lower body of the human body in different regions, and obtain the number of pollution points and pollution area in each region. The time-series correlation module is connected to the AI ​​video acquisition module and the fluorescent pollution automatic analysis module, respectively. It is used to match the timestamp of each violation action collected by the AI ​​video acquisition module with the pollution area, determine the effective pollution caused by the action and eliminate the false pollution remaining in the environment. The scoring module is used to determine the trainee's level based on the assessment's protection registration and historical assessment records, and to calculate the total assessment score.

[0021] Furthermore, the AI ​​video acquisition module includes at least three AI cameras and an edge computing box. It uses a key-point human posture model and a key-point hand model for action recognition, and uses a confidence-weighted Kalman filter fusion algorithm to output global human key-point coordinates. In this embodiment, the core polluting violations that can be identified and the quantitative judgment conditions are as follows: 1. The Euclidean distance between the key point of the hand and the key point of the outer surface of the protective clothing is less than 5cm and the duration is greater than 0.5s; 2. The wrist joint angle is greater than 150°, and the distance between the key point of the cuff and the key point of the skin is greater than 2cm; 3. The torso tilt angle is greater than 45°, and the distance between the key point of the hem of the protective clothing and the key point of the ground is less than 5cm; 4. The torso rotation angle is greater than 90° and the duration is greater than 1s; 5. The horizontal distance between the key points of the left and right feet is greater than 80cm.

[0022] Furthermore, the AI ​​video acquisition module uses a bidirectional LSTM network to analyze key point sequences and action temporal features, distinguishing between normal and illegal actions.

[0023] According to claim 1, the integrated training and assessment system for protective skills is characterized in that: the automatic fluorescence pollution analysis module specifically includes the following steps: First, Gaussian filtering for noise reduction and histogram equalization are used for image preprocessing. Secondly, the nested U-Net algorithm is used to segment the ultraviolet image into upper body and lower body regions; Secondly, a connected component algorithm is used to identify fluorescent residual regions and remove isolated noise points with an area smaller than a preset value. Finally, a kernel density estimation algorithm was used to generate a heat map of human body pollution. Specifically, the heat map uses color grading to represent the degree of pollution: blue for no pollution → green for light pollution → yellow for moderate pollution → red for heavy pollution, visually displaying high-risk pollution areas.

[0024] Furthermore, the time-series correlation module uses an Hidden Markov Model (HMM) to automatically divide the PPE removal process into eight time segments and establishes a mapping relationship between each time segment and the corresponding high-contamination-risk sub-areas within that time segment. These time segments include: removing goggles, removing outer gloves, removing protective clothing, removing inner gloves, removing shoe covers, removing hats, removing masks, and hand hygiene. Specifically: the goggles removal segment corresponds to the face and upper limbs; the outer gloves removal segment corresponds to the hands and wrists; the protective clothing removal segment corresponds to the torso and upper limbs; the inner gloves removal segment corresponds to the hands and wrists; the shoe covers removal segment corresponds to the lower limbs and ankles; the hat removal segment corresponds to the head and upper limbs; the mask removal segment corresponds to the face and upper limbs; and the hand hygiene segment corresponds to the hands.

[0025] Furthermore, the timing correlation module determines the effective pollution caused by the action and eliminates the false pollution remaining in the environment, specifically including the following steps: Step S1: Set a causal relationship time window centered on the occurrence time of each violation record; Step S2: Calculate the Euclidean distance between the left side of the contamination point to be determined and the operating area corresponding to the violation action; Step S3: Combine the temporal matching degree and spatial matching degree to calculate the confidence level of the causal association between the contamination point and the violation. Step S4: If the confidence level of causal association is greater than the preset value, the contamination point is determined to be a valid contamination caused by the action and is included in the deduction score; if the confidence level of causal association is less than the preset value, it is automatically removed.

[0026] Furthermore, the specific calculation formula for the scoring module is as follows: Final total score = Action standardization score × W1 + Pollution control score × W2 Wherein, W1+W2=1; the basic weights are set according to the protection level: Level 1 protection W1=0.6, W2=0.4, Level 2 protection W1=0.5, W2=0.5, Level 3 protection W1=0.4, W2=0.6.

[0027] Furthermore, the scoring module uses a clustering algorithm to automatically classify trainees into novice, proficient, and key personnel based on historical assessment data, and adjusts the coefficients by adding to the basic weights according to different levels.

[0028] The working principle is roughly as follows: The administrator creates assessment tasks in the system, sets the protection level, assessment time, and student list; the system automatically determines the student level based on the student's historical assessment records and calculates the dynamic weight of this assessment; the fluorescence pretreatment and imaging unit automatically prepares fluorescent tracers, sprays them evenly on the outer surface of the student's protective clothing, and allows them to air dry for at least 5 minutes.

[0029] Trainees log into the system to begin the assessment; the AI ​​video capture module simultaneously captures action videos, and the edge computing box performs real-time zoned action recognition; upon detecting a violation, the violation timestamp, region, and type are immediately recorded; and the action score for each region is calculated in real time. After the trainees complete the removal of PPE, the system starts the ultraviolet imaging camera to collect ultraviolet images and perform panoramic stitching; the fluorescence pollution automatic analysis module performs regional pollution analysis on the stitched images and generates pollution data and pollution heat maps for each region.

[0030] The time-series correlation module matches the time sequence of the violation with the polluted area, distinguishes between valid pollution and false pollution according to the judgment rules, eliminates false pollution and deducts points, and generates a table of correspondence between pollution-causing actions and pollution points.

[0031] The scoring module calculates the final assessment score according to preset rules; the system automatically generates an assessment report, which includes the trainee's basic information, zone scores, violation records, pollution analysis, pollution-causing actions, and personalized improvement suggestions.

[0032] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0033] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0034] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0035] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0036] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. An integrated training and assessment system for protective skills, characterized in that: AI video acquisition module, fluorescence preprocessing and imaging module, automatic fluorescence pollution analysis module, time series correlation module, and scoring module; The AI ​​video acquisition module is used to simultaneously acquire PPE donning and doffing action videos, divide human body movements into three regions—upper limbs, torso, and lower limbs—for regional recognition, add timestamps to each action and violation record, and calculate the standardization score of each region's actions. The fluorescence pretreatment and imaging module is used to prepare fluorescent tracers and spray them onto the outer surface of protective clothing, perform ultraviolet imaging on personnel after removal, and transmit the generated human ultraviolet image to the automatic fluorescence pollution analysis module. The automatic fluorescence pollution analysis module is used to receive ultraviolet images of the human body, perform pollution analysis on the upper and lower body of the human body in different regions, and obtain the number of pollution points and pollution area in each region. The time-series correlation module is connected to the AI ​​video acquisition module and the fluorescent pollution automatic analysis module, respectively. It is used to match the timestamp of each violation action collected by the AI ​​video acquisition module with the pollution area, determine the effective pollution caused by the action and eliminate the false pollution remaining in the environment. The scoring module is used to determine the trainee's level based on the assessment's protection registration and historical assessment records, and to calculate the total assessment score.

2. The integrated training and assessment system for protective skills according to claim 1, characterized in that: The AI ​​video acquisition module includes at least three AI cameras and an edge computing box. It uses a key point human pose model and a key point hand model for action recognition, and uses a confidence-weighted Kalman filter fusion algorithm to output global human key point coordinates.

3. The integrated training and assessment system for protective skills according to claim 1, characterized in that: The AI ​​video acquisition module uses a bidirectional LSTM network to capture key point sequences, analyze the temporal characteristics of actions, and distinguish between normal and illegal actions.

4. The integrated training and assessment system for protective skills according to claim 1, characterized in that: The automatic fluorescence contamination analysis module specifically includes the following steps: First, Gaussian filtering for noise reduction and histogram equalization are used for image preprocessing. Secondly, the nested U-Net algorithm is used to segment the ultraviolet image into upper body and lower body regions; Secondly, a connected component algorithm is used to identify fluorescent residual regions and remove isolated noise points with an area smaller than a preset value. Finally, a kernel density estimation algorithm was used to generate a heat map of human body pollution.

5. The integrated training and assessment system for protective skills according to claim 1, characterized in that: The time-series correlation module uses the HMM model to automatically divide the PPE removal process into eight time-series segments and establishes a mapping relationship between the time-series segments and the corresponding high-pollution-risk separate areas within the time-series segments. The time sequence includes removing goggles, outer gloves, protective clothing, inner gloves, shoe covers, hats, masks, and hand hygiene.

6. The integrated training and assessment system for protective skills according to claim 1, characterized in that: The time-series correlation module determines the effective pollution caused by the action and eliminates the false pollution remaining in the environment, specifically including the following steps: Step S1: Set a causal relationship time window centered on the occurrence time of each violation record; Step S2: Calculate the Euclidean distance between the left side of the contamination point to be determined and the operating area corresponding to the violation action; Step S3: Combine the temporal matching degree and spatial matching degree to calculate the confidence level of the causal association between the contamination point and the violation. Step S4: When the confidence level of causal relationship is greater than the preset value, the contamination point is determined to be a valid contamination caused by the action and is included in the deduction score; If the confidence level of causal association is less than the preset value, it will be automatically removed.

7. The integrated training and assessment system for protective skills according to claim 1, characterized in that: The specific calculation formula for the scoring module is as follows: Final total score = Action standardization score × W1 + Pollution control score × W2 Wherein, W1+W2=1; the basic weights are set according to the protection level: Level 1 protection W1=0.6, W2=0.4, Level 2 protection W1=0.5, W2=0.5, Level 3 protection W1=0.4, W2=0.

6.

8. The integrated training and assessment system for protective skills according to claim 1, characterized in that: The scoring module uses a clustering algorithm to automatically classify trainees into different levels based on historical assessment data, and adjusts the coefficients by adding weights to the base weights for different levels.

Images