Sterilization supply center article storage and automatic classification method based on intelligent algorithm

Abstract

The invention belongs to the technical field of article classification of disinfection and supply centers, and relates to an article storage and automatic classification method of a disinfection and supply center based on an intelligent algorithm. According to the method, firstly, an image recognition technology and an RFID tag technology are utilized to quickly recognize articles, and images and departments are recognized; related information of the current case is analyzed in combination with electronic medical record information, and whether an infected case exists or not is determined; constructing a risk classification model to further analyze names, material types and contact types of the articles, and performing feature fusion with department and infection case information to realize accurate classification of the articles; according to the technical scheme, the disinfection standard database is matched, the appropriate disinfection process is determined for each type of articles, the problem that the disinfection processes are not matched is solved, finally, the articles of different types are stored in different storage areas, follow-up disinfection and distribution are facilitated, and the working efficiency, accuracy and safety of the disinfection supply center are overall improved.

Description

Technical Field

[0001] The present application belongs to the technical field of item classification in a disinfection supply center, and more specifically, to an item storage and automatic classification method in a disinfection supply center based on an intelligent algorithm. Background Art

[0002] The disinfection and supply center is a vital department in a medical institution. It is mainly responsible for processing and supplying the medical supplies needed by various departments of the hospital. These items need to undergo a strict cleaning, disinfection, packaging and sterilization process before use to ensure that they will not cause cross infection to patients during use.

[0003] The current steps for the Sterilization Supply Center to process recycled items are as follows:

[0004] Recycling: Used instruments and articles should be promptly placed in a sealed waste box to avoid drying, and the name and quantity of the package should be recorded on the box cover; the supply room staff will go to the clinical department twice a day to recycle the dirty instruments and articles, and hand over and record with the clinical department staff; after recycling, the recycling staff will hand over the number of articles to the cleaning staff, and count and check whether the articles in the package are complete; after each recycling, the recycling box should be cleaned and disinfected, and stored in a dry place; used disposable articles and medical waste shall not be recycled to the disinfection supply center for re-operation and treatment;

[0005] Classification: Classify the equipment according to its different materials, properties and contamination conditions;

[0006] Cleaning and disinfection: Different cleaning and disinfection methods are used for different instruments and items; the basic process of cleaning and disinfection is as follows:

[0007] Pre-wash (tap water) - cleaning (manual + enzyme) - rinsing (tap water) - disinfection (wet heat disinfection) - final rinse (deionized water or distilled water) - drying - equipment inspection and maintenance - packaging - sterilization - storage - distribution;

[0008] However, at present, the classification is done by manual understanding of the objects, which will result in the following situations:

[0009] Poor classification accuracy: Since the classification is done based on manual knowledge of the items, it is easily affected by personal factors, resulting in inaccurate classification results, which may mix infected items with uninfected items, increasing the risk of cross-infection;

[0010] Mismatched disinfection procedures: Since it is impossible to accurately identify the infection cases of items, the selection of disinfection procedures is not accurate enough, which may cause insufficient disinfection of items. Summary of the invention

[0011] The present invention provides an article storage and automatic classification method for a disinfection supply center based on an intelligent algorithm, and is intended to solve the technical problems of poor classification accuracy and mismatch of disinfection processes in the classification method currently used in disinfection supply centers.

[0012] The method for storing and automatically classifying items in a disinfection supply center based on an intelligent algorithm includes the following steps:

[0013] Step 1: Use image recognition technology to collect images of items through cameras installed at each entrance and sorting area of ​​the disinfection supply center; and identify the department mentioned in the item based on RFID tag technology;

[0014] Step 2: Based on the electronic medical record information and the identified department, obtain the relevant information of the current case of the corresponding department, extract whether there is an infection case within the predetermined time based on the relevant information of the current case, obtain the label of whether there is an infection case in the corresponding department, and extract the name of the infection case if there is an infection case;

[0015] Step 3: Based on the collected images, the labels of the departments and whether there are any infection cases in the corresponding departments, the collected images are input into the constructed risk classification model, and the collected images are processed based on the risk classification model to obtain the names, material types and contact types of the objects. The results obtained based on image processing are then fused with the labels of the departments and whether there are any infection cases in the corresponding departments, and the objects are classified to obtain the risk classification results.

[0016] Step 4: Based on the name of the infected case, the material type of the item, and the risk classification results, match them with the pre-built disinfection standard database to determine the disinfection process corresponding to the risk classification results;

[0017] Step 5: According to the classification results, sort the items with the same risk classification results into corresponding storage areas and prepare to execute the corresponding disinfection process.

[0018] The present invention firstly utilizes image recognition technology and RFID tag technology to quickly identify objects, images and departments; then, it analyzes the relevant information of the current case in combination with electronic medical record information to determine whether there is an infected case, thereby providing an important decision-making basis for the subsequent disinfection process; then, it constructs a risk classification model to further analyze the name, material type and contact type of the object, and integrates the features with the department and infected case information to achieve accurate classification of the objects, thereby ensuring the separation of infected objects and non-infected objects and reducing the risk of cross infection; then, by matching the disinfection standard database, it determines the appropriate disinfection process for each type of object, thereby solving the problem of mismatching disinfection processes and ensuring the disinfection effect; finally, it stores different types of objects in different storage areas, thereby enabling objects of the same risk classification to be stored in an orderly manner, facilitating subsequent disinfection and distribution, thereby improving the work efficiency, accuracy and safety of the disinfection supply center as a whole.

[0019] Preferably, the step of extracting whether there is a case infection within a predetermined time based on the relevant information of the current case is as follows:

[0020] Timestamp acquisition: obtain the timestamp for identifying the department described in the item based on RFID tag technology;

[0021] Determine time: take the day to which the identified timestamp belongs as the predetermined time;

[0022] Extract whether there are infected cases: Extract the information of all cases on that day, and determine whether there are infected cases among all cases on that day. If yes, mark it as 1, if not, mark it as 0.

[0023] Preferably, the specific structure of the risk classification model is as follows:

[0024] Input layer: Input the object images captured by the camera, the departments, and the infection risk labels extracted from the case information within the departments;

[0025] The image processing module includes an object name recognition unit, a material type recognition unit, and a contact type recognition unit;

[0026] The object name recognition unit recognizes the object image through a CNN network and outputs the name of the object;

[0027] The material type identification unit extracts the material information of the object in the object image based on the CNN network;

[0028] The contact type identification unit infers the contact type of the object according to the name and material type of the object in combination with predefined rules;

[0029] Feature fusion layer: Through weighted feature fusion, the name of the object, material type, contact type, department information, and infection case label are weighted and spliced ​​to obtain the fused features;

[0030] Risk classification module: Classify items based on fusion features and multi-layer perceptrons to obtain risk values, and then divide items into corresponding risk levels based on the risk values.

[0031] Preferably, the item name recognition unit uses ResNet50 as the basic structure, and combines the attention mechanism and multi-scale feature fusion to extract the item name. The structure is as follows:

[0032] Input layer: input the image obtained by the camera;

[0033] Convolutional layer: The pre-trained ResNet50 network model is used for feature extraction, where the ResNet50 network model includes multiple convolutional layers and multiple residual modules. Feature extraction is performed based on multiple convolutional layers and multiple residual modules to obtain a high-dimensional feature representation of the image;

[0034] Attention mechanism module: After the last convolutional layer of the ResNet50 network model, a Squeeze-and-Excitation module is added. After the input of the last convolutional layer enters the Squeeze-and-Excitation module, global average pooling is first performed to convert the feature map of each channel into a scalar value through the average pooling operation. Then, the weight coefficient of the channel is generated based on the fully connected layer. Finally, each channel of the feature map is multiplied by the corresponding weight coefficient to strengthen the focus on key information.

[0035] Fully connected layer: fully connects the features output by the attention mechanism module and finally outputs the name of the identified object;

[0036] Output layer: Based on a Softmax classification layer, it is used to classify different item names.

[0037] Preferably, the CNN network in the material type recognition unit includes an input layer, a first convolutional layer, a second convolutional layer, a third convolutional layer, a fully connected layer and an output layer;

[0038] The first convolution layer receives the data of the input layer, uses a 3×3 convolution kernel, an output channel of 64, a step size of 1, padding of the same, and an activation function of ReLU; wherein the first convolution layer is connected to the first maximum pooling layer, and the pooling size of the first maximum pooling layer is 2×2, and the step size is 2;

[0039] The second convolution layer receives the output of the first maximum pooling layer, uses a 3×3 convolution kernel, an output channel of 128, a step size of 1, padding of the same, and an activation function of ReLU; wherein the second convolution layer is connected to the second maximum pooling layer, and the pooling size of the second maximum pooling layer is 2×2, and the step size is 2;

[0040] The third convolution layer receives the output of the second maximum pooling layer, uses a 3×3 convolution kernel, an output channel of 256, a step size of 1, padding of the same, and an activation function of ReLU; wherein the third convolution layer is connected to the third maximum pooling layer, and the pooling size of the third maximum pooling layer is 2×2, and the step size is 2;

[0041] Fully connected layer: First, the feature map output by the third pooling layer is flattened. The flattened feature map is input into the fully connected layer for processing and outputting the predicted value of the category. The fully connected layer includes 128 neurons, uses the ReLU activation function, and sets a dropout rate of 0.5 to prevent overfitting.

[0042] Output layer: The softmax function is used to convert the output of the fully connected layer into a probability distribution, and the category with the highest probability is used as the prediction result of the model.

[0043] Preferably, the specific steps of the feature fusion layer are as follows:

[0044] Assign a learnable weight to each input feature, and then multiply the weight by the corresponding feature to obtain the weighted feature;

[0045] Each weighted feature is concatenated to form a new feature vector, and the new feature vector is input as a fusion feature into the multi-layer perceptron for risk classification.

[0046] Preferably, the multilayer perceptron comprises an input layer, a first hidden layer, a second hidden layer and an output layer;

[0047] The input layer is used to receive fusion features;

[0048] The first hidden layer is used to process the fusion features input in the input layer, wherein the first hidden layer includes n neurons and adopts a ReLU activation function;

[0049] The second hidden layer is used to process the data output by the first hidden layer, wherein the second hidden layer comprises m neurons, where m is greater than n, and a ReLU activation function is used;

[0050] The output layer is composed of one neuron, which outputs the risk value of the item. The output value is processed by the Sigmoid activation function and limited to the range of [0,1].

[0051] Preferably, step 4 comprises the following steps:

[0052] Based on the material type of the article and the result of risk classification, a predefined risk level disinfection mapping table is queried, wherein the risk level disinfection mapping table has disinfection processes corresponding to different disinfection levels, wherein the disinfection process includes the name of the disinfection process and the disinfection method; wherein the risk level disinfection mapping table constructs a risk level disinfection mapping table for each material type, and when querying, firstly obtain the risk level disinfection mapping table of the corresponding material type based on the material type, and then query the corresponding disinfection process in the table based on the risk classification result;

[0053] Finally, the disinfection process of the query is modified based on the name of the infected case.

[0054] Preferably, the specific steps of correcting the queried disinfection process based on the name of the infection case are as follows:

[0055] Based on the name of the infection case, query the infection case disinfection mapping table to determine the specific disinfection method required for the infection case; wherein the infection case mapping table contains a mapping relationship between the infection case, the step, and the disinfection method;

[0056] Based on the steps and disinfection methods in the table, determine whether the corresponding steps in the disinfection process corresponding to the risk classification result query table use the same disinfection method. If the same disinfection method is used, the disinfection process corresponding to the risk classification query table is used as the output result; if different, the disinfection method involved in the infection case mapping table replaces the disinfection method corresponding to the steps involved in the risk classification query table, and the disinfection process after the replacement of the disinfection method is output.

[0057] The beneficial effects of the present invention include:

[0058] The present invention firstly utilizes image recognition technology and RFID tag technology to quickly identify objects, images and departments; then, it analyzes the relevant information of the current case in combination with electronic medical record information to determine whether there is an infected case, thereby providing an important decision-making basis for the subsequent disinfection process; then, it constructs a risk classification model to further analyze the name, material type and contact type of the object, and integrates the features with the department and infected case information to achieve accurate classification of the objects, thereby ensuring the separation of infected objects and non-infected objects and reducing the risk of cross infection; then, by matching the disinfection standard database, it determines the appropriate disinfection process for each type of object, thereby solving the problem of mismatching disinfection processes and ensuring the disinfection effect; finally, it stores different types of objects in different storage areas, thereby enabling objects of the same risk classification to be stored in an orderly manner, facilitating subsequent disinfection and distribution, thereby improving the work efficiency, accuracy and safety of the disinfection supply center as a whole. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0060] Figure 1 An overall step block diagram provided for an embodiment of the present invention. Figure 2 A schematic diagram of the structure of a risk classification model provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0061] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0062] See also Figure 1 As shown, the method for storing and automatically classifying items in a disinfection supply center based on an intelligent algorithm includes the following steps:

[0063] Step 1: Use image recognition technology to collect images of items through cameras installed at each entrance and sorting area of ​​the disinfection supply center; and identify the department mentioned in the item based on RFID tag technology;

[0064] As mentioned in the background technology, used instruments and item packages should be promptly placed in a waste bin for sealed storage. At this time, we can identify the corresponding department through RFID tag technology. We only need to set the corresponding department label on one side of the waste bin, and then scan the code with a scanner to obtain the corresponding department name. As for image recognition technology, we can install a camera above the sorting center. When it is necessary to open the lid and sort the items in the waste bin, we use the camera to collect images of the inside of the waste bin.

[0065] Step 2: Based on the electronic medical record information and the identified department, obtain the relevant information of the current case of the corresponding department, extract whether there is an infection case within the predetermined time based on the relevant information of the current case, obtain the label of whether there is an infection case in the corresponding department, and extract the name of the infection case if there is an infection case;

[0066] The steps of extracting whether there is a case infection within a predetermined time based on the relevant information of the current case are as follows:

[0067] Timestamp acquisition: obtain the timestamp for identifying the department described in the item based on RFID tag technology;

[0068] Determine time: take the day to which the identified timestamp belongs as the predetermined time;

[0069] Extract whether there are infected cases: Extract the information of all cases on that day, and determine whether there are infected cases among all cases on that day. If yes, mark it as 1, if not, mark it as 0.

[0070] As mentioned in the background technology, medical supplies need to be recycled twice a day in the corresponding department; and by obtaining the timestamp when the code is scanned, the case status of the department on the date of the timestamp is obtained. If there is an infection case, it is marked and the name of the infection case is obtained; and then the infection case is taken into account;

[0071] Step 3: Based on the collected images, the labels of the departments and whether there are any infection cases in the corresponding departments, the collected images are input into the constructed risk classification model, and the collected images are processed based on the risk classification model to obtain the names, material types and contact types of the objects. The results obtained based on image processing are then fused with the labels of the departments and whether there are any infection cases in the corresponding departments, and the objects are classified to obtain the risk classification results.

[0072] See also Figure 2 As shown, the specific structure of the risk classification model is as follows:

[0073] Input layer: Input the object images collected by the camera, the infection risk labels extracted from the departments and the case information within the departments; the input layer includes two input units, one input unit inputs the object images, and the other unit inputs the infection risk labels extracted from the departments and the case information within the departments; the object images enter the image processing module, and the infection risk labels extracted from the departments and the case information within the departments enter the feature fusion layer;

[0074] The image processing module includes an object name recognition unit, a material type recognition unit, and a contact type recognition unit;

[0075] The object name recognition unit recognizes the object image through a CNN network and outputs the name of the object;

[0076] The object name recognition unit uses ResNet50 as the basic structure, in which the pre-trained weights are trained on datasets such as ImageNet, and the object names are extracted by combining the attention mechanism and multi-scale feature fusion. The structure is as follows:

[0077] Input layer: Input the image H×W×C acquired by the camera, where H represents the height of the image; W represents the width of the image; and C represents the number of channels. Before inputting the image into the input layer, the image needs to be normalized to ensure that the value of each pixel is within a reasonable range, such as mapping the pixel value of the image from 0-255 to 0-1. The size of the input image is scaled or cropped to make the image meet the size requirements of the input layer (e.g., 224×224).

[0078] Convolutional layer: The pre-trained ResNet50 network model is used for feature extraction, wherein the ResNet50 network model includes multiple convolutional layers and multiple residual modules. Feature extraction is performed based on multiple convolutional layers and multiple residual modules to obtain a high-dimensional feature representation of the image; wherein the image is subjected to feature extraction through multiple convolutional layers, and the convolutional layers extract local information of the image (such as low-level features such as edges, textures, and colors). The residual module uses skip connections to avoid small gradients and ensure that features can be effectively propagated in deeper networks. The residual module outputs a deep feature map containing high-level information from the original image (such as object shape, texture, etc.); exemplary:

[0079] Suppose we have an initial input image I0 (size is 224×224×3), and after a series of convolutional layers and residual modules, we get a deep feature map.

[0080] The first residual module (Block1):

[0081] Input feature map size: 224×224×64;

[0082] Convolution operation: extract low-level features such as edges and textures;

[0083] Skip connection: add the input feature map to the convolution output;

[0084] Output feature map: 224×224×64;

[0085] Second residual module (Block2):

[0086] Input feature map: 224×224×64;

[0087] Convolution operation: further extract more complex patterns, local object features, etc.

[0088] Output feature map: 224×224×64;

[0089] Continue similar processing until the 49th residual module (Block49);

[0090] The last convolutional layer:

[0091] The last convolutional layer of ResNet50 outputs a deep feature map with a size of 7×7×2048; the feature map contains high-level information of the image, such as object shape, texture, position, etc. This feature map will be used as the input of the attention mechanism module.

[0092] Attention mechanism module: After the last convolutional layer of the ResNet50 network model, a Squeeze-and-Excitation module is added. After the input of the last convolutional layer enters the Squeeze-and-Excitation module, global average pooling is first performed to convert the feature map of each channel into a scalar value through the average pooling operation. Then, the weight coefficient of the channel is generated based on the fully connected layer. Finally, each channel of the feature map is multiplied by the corresponding weight coefficient to strengthen the focus on key information.

[0093] Exemplary:

[0094] Global Average Pooling: Global Average Pooling is performed on each channel, compressing each 7×7 feature map into a single scalar (average value). The output is a 2048-dimensional vector representing the global features of each channel:

[0095]

[0096] Where: x i,j,c represents the feature value of channel c at position (i, j) in the feature map; z c represents the global average pooling value of channel c;

[0097] Fully connected layer generates channel weights: The 2048-dimensional vector obtained by global average pooling is sent to a fully connected layer. This layer generates the weight coefficient of each channel through the activation function. Specifically, the SE module generates the channel weight coefficient through two fully connected layers (FC1 and FC2):

[0098] w c =σ(W2·δ(W1·z));

[0099] Where: W1 and W2 represent the weight matrices of the fully connected layer; σ represents the Sigmoid activation function; δ represents the ReLU activation function;

[0100] Weighted feature map: By multiplying the feature map of each channel by the corresponding weight coefficient w c , to enhance important features. The weighted feature map can highlight the key information that needs to be paid attention to in object name recognition.

[0101] Output: The weighted feature map still has a size of 7×7×2048, but the features of each channel have been weighted according to the attention mechanism;

[0102] Fully connected layer: Perform a fully connected operation on the features output by the attention mechanism module, and finally output the name of the identified object. The specific processing steps are as follows:

[0103] Flatten: Flatten the 7×7×2048 feature map into a one-dimensional vector with a size of 7×7×2048=100352-dimensional vector; input the flattened feature vector into the fully connected layer, map the flattened feature vector (100352 dimensions) to the category space, that is, generate a vector containing category probabilities; the fully connected layer realizes the mapping from input to output through the learned weight matrix; each output neuron is the weighted sum of the input vector and a set of weight matrices, plus a bias term; output an N-dimensional vector, representing the prediction score of each category;

[0104] Output layer: Based on a Softmax classification layer, it is used to classify different item names; the output of the network is converted into a probability distribution through the Softmax classification layer (Softmax function). Each element of this probability distribution is between [0,1], and the sum of all elements is 1; the output of the Softmax function is the predicted probability of each category, and the network will eventually select the category with the largest probability as the final prediction result.

[0105] The loss function of the object name recognition unit adopts the cross entropy loss function to calculate the difference between the true label and the probability of the model output; the loss value is used for back propagation to update the model parameters.

[0106] The material type identification unit extracts the material information of the object in the object image based on the CNN network;

[0107] The CNN network in the material type recognition unit includes an input layer, a first convolutional layer, a second convolutional layer, a third convolutional layer, a fully connected layer and an output layer;

[0108] The first convolution layer receives the data of the input layer, uses a 3×3 convolution kernel, an output channel of 64, a step size of 1, padding of the same, and an activation function of ReLU; wherein the first convolution layer is connected to the first maximum pooling layer, and the pooling size of the first maximum pooling layer is 2×2, and the step size is 2;

[0109] The second convolution layer receives the output of the first maximum pooling layer, uses a 3×3 convolution kernel, an output channel of 128, a step size of 1, padding of the same, and an activation function of ReLU; wherein the second convolution layer is connected to the second maximum pooling layer, and the pooling size of the second maximum pooling layer is 2×2, and the step size is 2;

[0110] The third convolution layer receives the output of the second maximum pooling layer, uses a 3×3 convolution kernel, an output channel of 256, a step size of 1, padding of the same, and an activation function of ReLU; wherein the third convolution layer is connected to the third maximum pooling layer, and the pooling size of the third maximum pooling layer is 2×2, and the step size is 2;

[0111] Fully connected layer: First, the feature map output by the third pooling layer is flattened. The flattened feature map is input into the fully connected layer for processing and outputting the predicted value of the category. The fully connected layer includes 128 neurons, uses the ReLU activation function, and sets a dropout rate of 0.5 to prevent overfitting.

[0112] Output layer: The softmax function is used to convert the output of the fully connected layer into a probability distribution, and the category with the highest probability is used as the prediction result of the model.

[0113] The contact type identification unit infers the contact type of the object according to the name and material type of the object in combination with predefined rules;

[0114] For example: contact types include direct contact, invasive contact, epidermal contact, and indirect contact;

[0115] Among them, direct contact is defined as surgical instruments, medical devices or any objects directly contacting the patient's skin, mucous membranes or tissues.

[0116] Invasive contact is defined as emphasizing the invasiveness of surgery or medical procedures, which means that surgical instruments or medical tools enter the patient's body and come into contact with internal tissues or organs.

[0117] Epidermal contact is contact with the patient's superficial skin;

[0118] Indirect contact refers to indirect contact between an object and a patient or the environment, usually when an object transmits pathogens by contacting the environment or surface.

[0119] For example, the object type we identified is a scalpel, and a scalpel is generally used to cut skin diseases deep into the subcutaneous tissue, muscle layer or even deeper. Therefore, based on this logic, a scalpel is an invasive contact.

[0120] Another example is bedding, which directly contacts the epidermis, so it is epidermal contact.

[0121] We determine the contact type of the corresponding object by pre-defining the above rules. Secondly, we can build a mapping table, which contains the mapping of object names and contact types. As mentioned in the above rules, a scalpel corresponds to invasive contact. Given a mapping relationship, we can query the contact relationship of the corresponding object through the mapping table.

[0122] Feature fusion layer: Through weighted feature fusion, the name of the object, material type, contact type, department information, and infection case label are weighted and spliced ​​to obtain the fused features;

[0123] The specific steps of the feature fusion layer are as follows:

[0124] Assign a learnable weight to each input feature, and then multiply the weight by the corresponding feature to obtain the weighted feature;

[0125] Each weighted feature is concatenated to form a new feature vector, and the new feature vector is input as a fusion feature into the multi-layer perceptron for risk classification.

[0126] Based on the above, we can know that the input of our risk classification module includes weighted item names, material types, contact types, department information and infection case labels; the level of risk is most correlated with contact types and infection case labels, so a higher initial weight can be assigned, and a lower weight is assigned to department information, item names and material types; the sum of the weights corresponding to item names, material types, contact types, department information and infection case labels is 1; for example, the initial weights of contact types and infection case labels are both 0.35; the weights of the other three input features are all 0.1.

[0127] Risk classification module: classifies items based on fusion features and multi-layer perceptron to obtain risk values, and then divides items into corresponding risk levels based on the risk values;

[0128] The multilayer perceptron includes an input layer, a first hidden layer, a second hidden layer and an output layer;

[0129] The input layer is used to receive fusion features;

[0130] The first hidden layer is used to process the fusion features input in the input layer, wherein the first hidden layer includes n neurons and adopts a ReLU activation function;

[0131] The second hidden layer is used to process the data output by the first hidden layer, wherein the second hidden layer comprises m neurons, where m is greater than n, and a ReLU activation function is used;

[0132] The output layer is composed of one neuron, which outputs the risk value of the item. The output value is processed by the Sigmoid activation function and limited to the range of [0,1].

[0133] In this embodiment, since the data dimension used is small, the multi-layer perceptron does not use the Dropout layer, as follows:

[0134] The input features are as follows:

[0135] Item name (text feature): If you use pre-trained word embeddings (such as Word2Vec or GloVe), each item name will be mapped to a fixed-dimensional vector (32 or 64 dimensions).

[0136] Material type (discrete category): usually one-hot encoding is used. Assuming there are 6 material types, the dimension is 6.

[0137] Contact type (discrete category): If it is encoded by one-hot, assuming there are 5 contact types, the dimension is 5.

[0138] Department information (discrete categories): Department information is processed through an embedding layer, assuming the embedding dimension is 6.

[0139] Infection case label (binary variable): This is a scalar with value 0 or 1 and dimension 1.

[0140] Therefore, after embedding and encoding, the input dimension will be the sum of the following parts:

[0141] Item names: 64 (assuming 64-dimensional embedding)

[0142] Material Type: 6

[0143] Contact Type: 5

[0144] Department Information: 6

[0145] Infection case label: 1

[0146] Therefore, the total dimension of the input features is: 64+6+5+6+1=82. This dimension is not particularly high for a neural network. Compared with some other complex tasks, the 82-dimensional input does not make the network overly complex, nor does it need to rely on Dropout to prevent overfitting. Secondly, due to weighted feature fusion, by assigning learnable weights to each input feature, the model can adaptively adjust the importance of features according to the training data. Weighted feature fusion makes the contribution of each feature clearer, thereby helping the model to be more efficient in processing features. This method reduces the model's over-reliance on all features to a certain extent, indirectly preventing overfitting.

[0147] In this embodiment, the department information is processed through the embedding layer and converted into a dense low-dimensional vector, which can better capture the relationship between the department and the item risk. Different departments have different infection risks. For example, surgery may involve a higher risk of external infection, while departments such as ICU may face a higher risk of cross infection. In this way, the model can automatically adjust the risk assessment of items according to department information and provide more accurate risk classification.

[0148] Secondly, the consideration of contact type reflects the way in which the item contacts the human body or other surfaces, which involves the path of infectious disease transmission. In medical and public health environments, the contact type of an item has a great impact on the risk of infection. Through One-Hot encoding, different contact types are converted into discrete features to help the model understand the different effects of different types of contact on the risk of items. The infection case label directly indicates whether the item is associated with an infection case, which is crucial for the risk assessment of the item. As a binary feature (0 or 1), it clearly expresses the infection association of the item, helping the model to quickly and accurately identify infected items.

[0149] However, in the prior art, whether there are infection cases in the department and whether the corresponding disinfection process needs to be carried out are all known through manual notification. The present invention directly obtains whether there are infection cases in the corresponding department through electronic medical records, and predicts the risk type based on the obtained information, so that the final prediction result is fully considered and the technical problem of forgetting in the prior art is avoided. Although disinfection monitoring will be carried out after disinfection to prevent items with pathogens from being released from the warehouse, it will also cause duplication of workload. Therefore, we effectively avoid the problem of duplication of workload; and the present invention is more intelligent and avoids the technical problems of poor classification accuracy and mismatch of disinfection processes caused by manual classification.

[0150] Step 4: Based on the name of the infected case, the material type of the item, and the risk classification results, match them with the pre-built disinfection standard database to determine the disinfection process corresponding to the risk classification results;

[0151] The step 4 comprises the following steps:

[0152] Based on the material type and risk classification results of the article, a predefined risk level disinfection mapping table is queried (see Table 2, which is a risk level disinfection mapping table for stainless steel materials), wherein the risk level disinfection mapping table is constructed with disinfection processes corresponding to different disinfection levels, wherein the disinfection process includes the name of the disinfection process and the disinfection method; wherein the risk level disinfection mapping table is constructed for each material type, and when querying, the risk level disinfection mapping table corresponding to the material type is first obtained based on the material type, and then the corresponding disinfection process in the table is queried based on the risk classification result;

[0153] Table 2: Risk level disinfection mapping table for stainless steel materials

[0154]

[0155]

[0156]

[0157] Based on Table 2 above, it can be seen that a risk level disinfection mapping table for stainless steel materials is given. We give a corresponding risk level disinfection mapping table for each material. For example, corresponding risk level disinfection mapping tables are given for glass, fabric, plastic and other materials respectively. Of course, the above examples are not limitations of the present invention. Regarding the level setting, we can set multiple levels, not just five levels. Specifically, different levels can be established according to different disinfection process standards in actual conditions.

[0158] Finally, the disinfection process of the query is modified based on the name of the infected case.

[0159] The specific steps of the disinfection process for modifying the query based on the name of the infected case are as follows:

[0160] Based on the name of the infection case, query the infection case disinfection mapping table to determine the specific disinfection method required for the infection case; wherein the infection case mapping table contains the mapping relationship between the infection case, the step, and the disinfection method, see Table 3;

[0161] Table 3. Infection case mapping table

[0162]

[0163]

[0164] Based on the steps and disinfection methods in the table, determine whether the corresponding steps in the disinfection process corresponding to the risk classification result query table use the same disinfection method. If the same disinfection method is used, the disinfection process corresponding to the risk classification query table is used as the output result; if different, the disinfection method involved in the infection case mapping table replaces the disinfection method corresponding to the step involved in the risk classification query table, and outputs the disinfection process after the replacement of the disinfection method;

[0165] For example, referring to Table 2, if we output the highest risk, that is, risk level 1, then the corresponding disinfection process in risk level 1 is output; if the name of the infected case is tuberculosis, then check whether the disinfection and sterilization method output in Table 2 meets the requirements of tuberculosis in Table 3. Based on Table 2, it can be seen that it does not meet the requirements, so it is necessary to replace the disinfection and sterilization method combined with the disease in Table 3 with the disinfection and sterilization method output in Table 2, and finally output the adjusted disinfection process.

[0166] Based on the above, different infection cases (such as tuberculosis, hepatitis, drug-resistant bacteria, etc.) have different requirements for disinfection methods. These requirements are usually closely related to factors such as the infectivity, drug resistance, and transmission route of the pathogen. By adjusting the disinfection process according to the specific type of infection case, the targetedness and effectiveness of the disinfection can be ensured. By adjusting the disinfection process based on the type of infection case, cross-contamination of different pathogens can be effectively avoided. For example, for airborne pathogens such as tuberculosis, measures such as moist heat disinfection and ultraviolet disinfection are required, while for blood-borne diseases such as hepatitis, vinyl gas disinfection may be more suitable. Such refined management can reduce the risk of cross-infection, especially in highly sensitive environments (such as operating rooms, ICUs, etc.).

[0167] Step 5: According to the classification results, sort the items with the same risk classification results into corresponding storage areas and prepare to execute the corresponding disinfection process.

[0168] The present invention firstly utilizes image recognition technology and RFID tag technology to quickly identify objects, images and departments; then, it analyzes the relevant information of the current case in combination with electronic medical record information to determine whether there is an infected case, thereby providing an important decision-making basis for the subsequent disinfection process; then, it constructs a risk classification model to further analyze the name, material type and contact type of the object, and integrates the features with the department and infected case information to achieve accurate classification of the objects, thereby ensuring the separation of infected objects and non-infected objects and reducing the risk of cross infection; then, by matching the disinfection standard database, it determines the appropriate disinfection process for each type of object, thereby solving the problem of mismatching disinfection processes and ensuring the disinfection effect; finally, it stores different types of objects in different storage areas, thereby enabling objects of the same risk classification to be stored in an orderly manner, facilitating subsequent disinfection and distribution, thereby improving the work efficiency, accuracy and safety of the disinfection supply center as a whole.

[0169] The above are only preferred embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. A method for storing and automatically classifying items in a disinfection supply center based on an intelligent algorithm, characterized in that: The following steps are involved: Step 1: Use image recognition technology to collect images of items through cameras installed at each entrance and sorting area of ​​the disinfection supply center; and identify the department mentioned in the item based on RFID tag technology; Step 2: Based on the electronic medical record information and the identified department, obtain the relevant information of the current case of the corresponding department, extract whether there is an infection case within the predetermined time based on the relevant information of the current case, obtain the label of whether there is an infection case in the corresponding department, and extract the name of the infection case if there is an infection case; Step 3: Based on the collected images, the labels of the departments and whether there are any infection cases in the corresponding departments, the collected images are input into the constructed risk classification model, and the collected images are processed based on the risk classification model to obtain the names, material types and contact types of the objects. The results obtained based on image processing are then fused with the labels of the departments and whether there are any infection cases in the corresponding departments, and the objects are classified to obtain the risk classification results. Step 4: Based on the name of the infected case, the material type of the item, and the risk classification results, match them with the pre-built disinfection standard database to determine the disinfection process corresponding to the risk classification results; Step 5: According to the classification results, sort the items with the same risk classification results into corresponding storage areas and prepare to execute the corresponding disinfection process.

2. The method for storing and automatically classifying items in a disinfection supply center based on an intelligent algorithm according to claim 1, characterized in that: The steps of extracting whether there is a case infection within a predetermined time based on the relevant information of the current case are as follows: Timestamp acquisition: obtain the timestamp for identifying the department described in the item based on RFID tag technology; Determine time: take the day to which the identified timestamp belongs as the predetermined time; Extract whether there are infected cases: Extract the information of all cases on that day, and determine whether there are infected cases among all cases on that day. If yes, mark it as 1, if not, mark it as 0.

3. The method for storing and automatically classifying items in a disinfection supply center based on an intelligent algorithm according to claim 1, characterized in that: The specific structure of the risk classification model is as follows: Input layer: Input the object images captured by the camera, the departments, and the infection risk labels extracted from the case information within the departments; The image processing module includes an object name recognition unit, a material type recognition unit, and a contact type recognition unit; The object name recognition unit recognizes the object image through a CNN network and outputs the name of the object; The material type identification unit extracts the material information of the object in the object image based on the CNN network; The contact type identification unit infers the contact type of the object according to the name and material type of the object in combination with predefined rules; Feature fusion layer: Through weighted feature fusion, the name of the object, material type, contact type, department information, and infection case label are weighted and spliced ​​to obtain the fused features; Risk classification module: Classify items based on fusion features and multi-layer perceptrons to obtain risk values, and then divide items into corresponding risk levels based on the risk values.

4. The method for storing and automatically classifying items in a disinfection supply center based on an intelligent algorithm according to claim 3 is characterized in that: The object name recognition unit uses ResNet50 as the basic structure, and combines the attention mechanism and multi-scale feature fusion to extract the object name. The structure is as follows: Input layer: input the image obtained by the camera; Convolutional layer: The pre-trained ResNet50 network model is used for feature extraction, where the ResNet50 network model includes multiple convolutional layers and multiple residual modules. Feature extraction is performed based on multiple convolutional layers and multiple residual modules to obtain a high-dimensional feature representation of the image; Attention mechanism module: After the last convolutional layer of the ResNet50 network model, a Squeeze-and-Excitation module is added. After the input of the last convolutional layer enters the Squeeze-and-Excitation module, global average pooling is first performed to convert the feature map of each channel into a scalar value through the average pooling operation. Then, the weight coefficient of the channel is generated based on the fully connected layer. Finally, each channel of the feature map is multiplied by the corresponding weight coefficient to strengthen the focus on key information. Fully connected layer: fully connects the features output by the attention mechanism module and finally outputs the name of the identified object; Output layer: Based on a Softmax classification layer, it is used to classify different item names.

5. The method for storing and automatically classifying items in a disinfection supply center based on an intelligent algorithm according to claim 3, characterized in that: The CNN network in the material type recognition unit includes an input layer, a first convolutional layer, a second convolutional layer, a third convolutional layer, a fully connected layer and an output layer; The first convolution layer receives the data of the input layer, uses a 3×3 convolution kernel, an output channel of 64, a step size of 1, padding of the same, and an activation function of ReLU; wherein the first convolution layer is connected to the first maximum pooling layer, and the pooling size of the first maximum pooling layer is 2×2, and the step size is 2; The second convolution layer receives the output of the first maximum pooling layer, uses a 3×3 convolution kernel, an output channel of 128, a step size of 1, padding of the same, and an activation function of ReLU; wherein the second convolution layer is connected to the second maximum pooling layer, and the pooling size of the second maximum pooling layer is 2×2, and the step size is 2; The third convolution layer receives the output of the second maximum pooling layer, uses a 3×3 convolution kernel, an output channel of 256, a step size of 1, padding of the same, and an activation function of ReLU; wherein the third convolution layer is connected to the third maximum pooling layer, and the pooling size of the third maximum pooling layer is 2×2, and the step size is 2; Fully connected layer: First, the feature map output by the third pooling layer is flattened. The flattened feature map is input into the fully connected layer for processing and outputting the predicted value of the category. The fully connected layer includes 128 neurons, uses the ReLU activation function, and sets a dropout rate of 0.5 to prevent overfitting. Output layer: The softmax function is used to convert the output of the fully connected layer into a probability distribution, and the category with the highest probability is used as the prediction result of the model.

6. The method for storing and automatically classifying items in a disinfection supply center based on an intelligent algorithm according to claim 3, characterized in that: The specific steps of the feature fusion layer are as follows: Assign a learnable weight to each input feature, and then multiply the weight by the corresponding feature to obtain the weighted feature; Each weighted feature is concatenated to form a new feature vector, and the new feature vector is input as a fusion feature into the multi-layer perceptron for risk classification.

7. The method for storing and automatically classifying items in a disinfection supply center based on an intelligent algorithm according to claim 3 is characterized in that: The multilayer perceptron includes an input layer, a first hidden layer, a second hidden layer and an output layer; The input layer is used to receive fusion features; The first hidden layer is used to process the fusion features input in the input layer, wherein the first hidden layer includes n neurons and adopts a ReLU activation function; The second hidden layer is used to process the data output by the first hidden layer, wherein the second hidden layer comprises m neurons, where m is greater than n, and a ReLU activation function is used; The output layer is composed of one neuron, which outputs the risk value of the item. The output value is processed by the Sigmoid activation function and limited to the range of [0,1].

8. The method for storing and automatically classifying items in a disinfection supply center based on an intelligent algorithm according to claim 1, characterized in that: The step 4 comprises the following steps: Based on the material type of the article and the result of risk classification, a predefined risk level disinfection mapping table is queried, wherein the risk level disinfection mapping table has disinfection processes corresponding to different disinfection levels, wherein the disinfection process includes the name of the disinfection process and the disinfection method; wherein the risk level disinfection mapping table constructs a risk level disinfection mapping table for each material type, and when querying, firstly obtain the risk level disinfection mapping table of the corresponding material type based on the material type, and then query the corresponding disinfection process in the table based on the risk classification result; Finally, the disinfection process of the query is modified based on the name of the infected case.

9. The method for storing and automatically classifying items in a disinfection supply center based on an intelligent algorithm according to claim 8, characterized in that: The specific steps of the disinfection process for modifying the query based on the name of the infected case are as follows: Based on the name of the infected case, the infected case disinfection mapping table is queried to determine the specific disinfection method required for the infected case; The infection case mapping table includes the mapping relationship between infection cases, steps, and disinfection methods; Based on the steps and disinfection methods in the table, determine whether the corresponding steps in the disinfection process corresponding to the risk classification result query table use the same disinfection method. If the same disinfection method is used, the disinfection process corresponding to the risk classification query table is used as the output result; if different, the disinfection method involved in the infection case mapping table replaces the disinfection method corresponding to the steps involved in the risk classification query table, and the disinfection process after the replacement of the disinfection method is output.

Images