Intelligent surgical instrument counting system and method

Abstract

The invention relates to an intelligent surgical instrument counting system and method. The method comprises the steps that surgical instrument images located in a surgical instrument frame are collected in batches; identifying morphological characteristics in the surgical instrument image through the surgical instrument identification model, and outputting surgical instrument information corresponding to the morphological characteristics; and counting the surgical instrument information and generating a corresponding surgical instrument counting report. The CNN recognition technology can be utilized to carry out one-time image collection on surgical instruments passing through the instrument frame, form recognition of a plurality of surgical instruments on the images is carried out through the CNN model, so that one-time surgical instrument recognition and counting are achieved, the surgical instruments of the current batch can be rapidly counted, a counting report is generated by the background and sent to the PDA, code pasting for counting is not needed, and the counting efficiency is improved. And the defect of code sticking and counting is avoided. And the surgical instruments can be counted in batches, and the operation and use efficiency is greatly improved.

Description

Technical Field

[0001] The present disclosure relates to the technical field of medical instruments, and in particular to an intelligent surgical instrument inventory system, a one-time surgical instrument inventory method, and an electronic device. Background Art

[0002] In the operating room, the nursing staff must count the surgical instruments before each operation. The specific process is: the surgical instruments are placed in a surgical instrument box. Each time the counting is done, two nurses are required to count the number of surgical instruments one by one. Therefore, this manual counting method not only takes up nursing resources, but also has low efficiency.

[0003] In addition, with the development of code scanning technology, a method of scanning codes (nursing terminals such as PDA) for inventory has emerged. However, this method first requires a QR code to be attached to the surgical instrument frame or each instrument in the surgical instrument frame, and then the surgical instrument information can be collected and counted by scanning the terminal. However, this method has the following problems: First, the early stage of labeling requires a lot of manpower, time and other costs, which will affect the efficiency of the transportation and use of surgical instruments; moreover, manual counting is prone to errors, resulting in the omission of some surgical instruments; Second, the labels on surgical instruments will cause medical impacts. For example, if the QR code is made of adhesive paper, there is a risk that the label code will fall off and affect the safety of the operation, or the label material will come into contact with the wound and cause contact infection. Especially if it is a high-temperature disinfection lamp, the impact will be aggravated.

[0004] Counting pure surgical instruments and putting them into use is the best and safest application solution, so there is an urgent need to optimize and improve this inventory service. Summary of the invention

[0005] In order to solve the above problems, the present application proposes an intelligent surgical instrument inventory system, a one-time surgical instrument inventory method and an electronic device.

[0006] On the one hand, the present application proposes an intelligent surgical instrument inventory system, comprising the following steps: S1. Based on the surgical instrument image set, train and build a surgical instrument recognition model; S2, batch acquisition of surgical instrument images located in the surgical instrument frame; S3, importing the surgical instrument image into the surgical instrument recognition model, identifying morphological features in the surgical instrument image through the surgical instrument recognition model, and outputting surgical instrument information corresponding to the morphological features; S4. Count the surgical instrument information and generate a corresponding surgical instrument inventory report.

[0007] As an optional implementation scheme of the present application, optionally, S1, based on the surgical instrument image set, training and constructing a surgical instrument recognition model, including: Collecting a number of surgical instrument images and forming the surgical instrument image set after image preprocessing; Performing feature engineering on the surgical instrument image set, and extracting morphological features of each surgical instrument image in the surgical instrument image set using CNN; The morphological features are marked as surgical instruments, where the surgical instrument information includes at least one of the following information: surgical instrument attribute information, supply chain information, ownership information, sterilization supply information or dynamic price information; After the labeling is completed, a feature set consisting of a number of labeled morphological features is obtained; Dividing the feature set into a training set and a validation set according to a preset ratio; Importing the training set into a preset CNN model, performing CNN model training, and generating the surgical instrument recognition model; The recognition performance of the surgical instrument recognition model is verified using the verification set: If the verification is successful, the surgical instrument recognition model is deployed on the nurse station backend server and put into use; Otherwise, repeat the training according to the above steps.

[0008] As an optional implementation of the present application, optionally, S2, batch acquisition of surgical instrument images located in the surgical instrument frame, includes: Using an image acquisition device, photograph a number of surgical instruments laid out in a surgical instrument frame, batch-acquire surgical instrument images containing each surgical instrument, and upload them to a backend server at the nurse station; Using a contour recognition algorithm to identify the image contours of each surgical instrument in the surgical instrument image; Marking the image contours where each surgical instrument is located according to the identified image contours, and extracting image slices of each surgical instrument in the marked area; The image slices of each surgical instrument are sequentially saved in the database of the nurse station backend server.

[0009] As an optional implementation scheme of the present application, optionally, S3, importing the surgical instrument image into the surgical instrument recognition model, identifying morphological features in the surgical instrument image by the surgical instrument recognition model, and outputting surgical instrument information corresponding to the morphological features, includes: Reading the image slices of each surgical instrument in sequence and inputting them into the surgical instrument recognition model; The morphological features in the image slices are identified by the surgical instrument recognition model, and the surgical instrument information marked on the morphological features is matched and output.

[0010] As an optional implementation scheme of the present application, optionally, S4, counting the surgical instrument information and generating a corresponding surgical instrument inventory report, includes: Collecting statistics of surgical instrument information corresponding to the image slices of each surgical instrument; Writing the surgical instrument information into a preset inventory report to generate the surgical instrument inventory report for the surgical instruments currently located in the surgical instrument frame; The surgical instrument inventory report is saved and sent to the PDA terminal of the on-site nurse.

[0011] On the other hand, the present application proposes an intelligent surgical instrument inventory system for implementing the above-mentioned surgical instrument one-time inventory method, comprising: An image acquisition device, used to batch-acquire images of surgical instruments in the surgical instrument frame and upload them to the nurse station backend server; The nurse station backend server is used to import the surgical instrument image into the surgical instrument recognition model, identify the morphological features in the surgical instrument image through the surgical instrument recognition model, and output surgical instrument information corresponding to the morphological features; and count the surgical instrument information and generate a corresponding surgical instrument inventory report; PDA terminal, used to log in to the nurse station backend server, receive and view the surgical instrument inventory report; The image acquisition device and the PDA terminal are respectively connected to the backend server of the nurse station for communication.

[0012] In another aspect, the present application further provides an electronic device, comprising: processor; a memory for storing processor-executable instructions; Wherein, the processor is configured to implement the one-time inventory method for surgical instruments when executing the executable instructions.

[0013] Technical effects of the present invention: The present invention trains and constructs a surgical instrument recognition model based on a surgical instrument image set; batch collects surgical instrument images located in a surgical instrument frame; imports the surgical instrument images into the surgical instrument recognition model, identifies the morphological features in the surgical instrument images through the surgical instrument recognition model, and outputs surgical instrument information corresponding to the morphological features; counts the surgical instrument information and generates a corresponding surgical instrument inventory report. The CNN recognition technology can be used to perform a one-time image acquisition of the surgical instruments passing through the instrument frame, and the morphological recognition of several surgical instruments on the image through the CNN model, thereby realizing a one-time surgical instrument recognition and inventory, and can quickly count the current batch of surgical instruments and generate an inventory report by the background and send it to the PDA, so that there is no need to stick a code for inventory, avoiding the disadvantages of sticking a code for inventory. In addition, the present invention can count surgical instruments in batches, greatly improving their operation and use efficiency.

[0014] Further features and aspects of the present disclosure will become apparent from the following detailed description of exemplary embodiments with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate exemplary embodiments, features, and aspects of the disclosure and, together with the description, serve to explain the principles of the disclosure.

[0016] Figure 1 It is shown as a schematic diagram of the implementation process of the present invention; Figure 2 It is shown as a model training flow chart of the present invention; Figure 3 Shown is a schematic diagram of the system composition structure of the present invention; Figure 4 It is a schematic diagram showing the application of the electronic device of the present invention. DETAILED DESCRIPTION

[0017] Various exemplary embodiments, features and aspects of the present disclosure will be described in detail below with reference to the accompanying drawings. The same reference numerals in the accompanying drawings represent elements with the same or similar functions. Although various aspects of the embodiments are shown in the accompanying drawings, the drawings are not necessarily drawn to scale unless otherwise specified.

[0018] The word “exemplary” is used exclusively herein to mean “serving as an example, example, or illustration.” Any embodiment described herein as “exemplary” is not necessarily to be construed as preferred or advantageous over other embodiments.

[0019] In addition, in order to better illustrate the present disclosure, numerous specific details are given in the following specific embodiments. It should be understood by those skilled in the art that the present disclosure can also be implemented without certain specific details. In some examples, means, components and circuits well known to those skilled in the art are not described in detail in order to highlight the main purpose of the present disclosure. Example 1

[0020] like Figure 1 As shown, on one hand, the present application proposes an intelligent surgical instrument inventory system, comprising the following steps: S1. Based on the surgical instrument image set, train and build a surgical instrument recognition model; S2, batch acquisition of surgical instrument images located in the surgical instrument frame; S3, importing the surgical instrument image into the surgical instrument recognition model, identifying morphological features in the surgical instrument image through the surgical instrument recognition model, and outputting surgical instrument information corresponding to the morphological features; S4. Count the surgical instrument information and generate a corresponding surgical instrument inventory report.

[0021] like Figure 3 As shown, the present invention uses CNN to perform feature engineering on the image set and train to build a surgical instrument recognition model, and deploys the surgical instrument recognition model on the nurse station background server (or other application servers). An image acquisition device such as a CCD camera or other camera equipment is used to perform one-time batch image acquisition on several surgical instruments in the surgical instrument frame. The acquired image contains several surgical instruments. The acquired surgical instrument images are uploaded to the background and the background model is used for image recognition to identify the morphological features of each surgical instrument in the image, so as to match and output the corresponding surgical instrument information. The background records the results of the image recognition, and the inventory information of each surgical instrument contained in this frame is obtained by statistics. The list information of this batch of surgical instruments is quickly obtained in batches and sent to the nurse's handheld PDA terminal to achieve rapid inventory. The inventory efficiency is greatly improved, and there is no need to stick codes or scan codes for identification, which greatly reduces the risk of surgical contamination.

[0022] The image acquisition device can communicate with the background through USB or wireless; the PDA can communicate with the background through the intranet or wireless such as 5G. Please configure the network communication according to the existing equipment.

[0023] like Figure 2 As shown, as an optional implementation scheme of the present application, optionally, S1, based on the surgical instrument image set, training and constructing a surgical instrument recognition model, including: Collecting a number of surgical instrument images and forming the surgical instrument image set after image preprocessing; Performing feature engineering on the surgical instrument image set, and extracting morphological features of each surgical instrument image in the surgical instrument image set using CNN; The morphological features are marked as surgical instruments, where the surgical instrument information includes at least one of the following information: surgical instrument attribute information, supply chain information, ownership information, sterilization supply information or dynamic price information; After the labeling is completed, a feature set consisting of a number of labeled morphological features is obtained; Dividing the feature set into a training set and a validation set according to a preset ratio; Importing the training set into a preset CNN model, performing CNN model training, and generating the surgical instrument recognition model; The recognition performance of the surgical instrument recognition model is verified using the verification set: If the verification is successful, the surgical instrument recognition model is deployed on the nurse station backend server and put into use; Otherwise, repeat the training according to the above steps.

[0024] Here, the images of different surgical instruments are prepared by the administrator. Please note that several surgical instruments cannot overlap, and they need to be laid out flat without overlap after being placed in the surgical instrument frame.

[0025] CNN can be well used for training and constructing image recognition models. The structures of CNN such as convolutional layers and network layers are not described in detail in this embodiment.

[0026] The present invention uses a convolutional neural network (CNN) to extract morphological features of an image set and construct a corresponding model, which mainly includes the following main stages: 1. Data Preparation ‌Data Collection‌: Collect image datasets containing target morphological features. The datasets should cover different morphological changes and have sufficient quantity and diversity.

[0027] ‌Data Preprocessing‌: Preprocess the images, including image resizing, normalization (scaling pixel values ​​to a specific range, such as [0, 1]), data augmentation (such as rotation, flipping, cropping, etc.), and label encoding (if it is a classification task).

[0028] 2. Build a CNN model ‌Select Architecture‌: Select a suitable CNN architecture, such as AlexNet, VGG, ResNet, Inception, etc., or customize the network structure according to your needs.

[0029] ‌Define the model‌: ‌Input Layer‌: defines the dimensions of the input image (e.g., 224x224x3).

[0030] ‌Convolutional layer‌: Use multiple convolution kernels to extract image features, usually in conjunction with activation functions (such as ReLU) and batch normalization.

[0031] ‌Pooling layer‌: Reduces the dimension of the feature map, such as Max Pooling or Average Pooling.

[0032] ‌Fully Connected Layer‌: After feature extraction, high-dimensional features are mapped to the output space (e.g., the number of categories in a classification task).

[0033] ‌Output layer‌: Set according to the task type, such as using the softmax activation function in classification tasks.

[0034] 3. Model Training ‌Define loss function‌: Choose an appropriate loss function based on the task, such as cross entropy loss (classification task) or mean squared error (regression task).

[0035] Optimizer selection: Select a suitable optimization algorithm, such as SGD, Adam, RMSprop, etc., and set parameters such as learning rate and momentum.

[0036] ‌Batch training‌: Divide the data into small batches and perform iterative training. Set a reasonable batch size and number of training rounds (epochs). Input the training set into the model to train and learn the morphological features.

[0037] ‌Validation and Evaluation‌: Use the validation set to monitor model performance and prevent overfitting. You can use the Early Stopping strategy.

[0038] 4. Feature extraction ‌Feature Extractor‌: The trained CNN model can be used as a feature extractor. Usually, the last fully connected layer (or output layer) is removed, and the features extracted by the convolutional layer and the pooling layer are retained.

[0039] ‌Feature vector‌: The input image is passed through the feature extractor to obtain a high-dimensional feature vector. These feature vectors can be used for subsequent tasks such as classification, clustering, retrieval, etc.

[0040] Here we mainly use CNN to extract the morphological features of each surgical instrument (it is assumed that each surgical instrument has different shapes and sizes) to construct a feature set.

[0041] Here, the feature set can be divided into a training set and a validation set in a ratio of 7:3.

[0042] 5. Model building and optimization (for specific tasks) ‌Use extracted features‌: ‌Classification task‌: Input the extracted features into a classifier (such as SVM, logistic regression, KNN) for classification.

[0043] Clustering task: Use clustering algorithms (such as K-means, DBSCAN) to perform cluster analysis on features.

[0044] ‌Other tasks‌: such as image retrieval, anomaly detection, etc., design models according to specific needs.

[0045] ‌Fine-tuning‌: If the target task is very different from the task of the pre-trained model, you can fine-tune some layers of the pre-trained model to better adapt to the new task.

[0046] 6. Model Evaluation and Deployment ‌Model evaluation‌: Evaluate the model performance on the test set, using appropriate evaluation indicators (such as accuracy, F1-score, AUC, etc.). Use the validation set to input the model to verify the recognition accuracy. The administrator will count the verification results and determine whether the verification is passed. If it is passed, the model will be deployed and applied.

[0047] ‌Model deployment‌: Deploy the trained model and feature extractor to actual applications, and consider model optimization (such as quantization and pruning) to improve operational efficiency.

[0048] The model file can be recorded on the background server, and the model can be installed and applied.

[0049] As an optional implementation of the present application, optionally, S2, batch acquisition of surgical instrument images located in the surgical instrument frame, includes: Using an image acquisition device, photograph a number of surgical instruments laid out in a surgical instrument frame, batch-acquire surgical instrument images containing each surgical instrument, and upload them to a backend server at the nurse station; Using a contour recognition algorithm to identify the image contours of each surgical instrument in the surgical instrument image; Marking the image contours where each surgical instrument is located according to the identified image contours, and extracting image slices of each surgical instrument in the marked area; The image slices of each surgical instrument are sequentially saved in the database of the nurse station backend server.

[0050] For several surgical instruments laid flat in the surgical instrument frame, the warehouse manager can prepare them in advance, for example, using an intelligent robot or a conveyor belt to transport each surgical instrument frame to the inventory position. A proximity sensor can be set at the corresponding inventory position. When the surgical instrument frame is transported and close to the contact sensor, the sensor signal is collected and sent to the controller, and the controller controls the image acquisition device to take images. The image acquisition device performs a one-time image acquisition of the surgical instruments in the surgical instrument frame passing by, and obtains several images in batches and uploads them to the background server.

[0051] The backend server first performs contour recognition on each surgical instrument contained in the image, thereby identifying the image contour position where each surgical instrument is located. The present invention can use the contour recognition algorithm to identify the image area where each surgical instrument is located in the image, and recognize and mark the contour of the area, so as to facilitate the subsequent slicing of the image according to the contour, obtain a number of image slices and save them, input each image slice into the model, and the model recognizes the morphological features and outputs the corresponding slice image features.

[0052] The contour recognition algorithm is applied as follows: Contour recognition algorithms are an important technology in the field of computer vision and image processing, which are used to detect and identify the contours of objects from images or videos. These algorithms can extract the edge information of objects in images, and then realize object recognition, tracking and analysis. Contour recognition has a wide range of applications in many fields such as robot navigation, video surveillance, medical imaging, and autonomous driving.

[0053] Basic steps of contour recognition 1. Image preprocessing: ‌Grayscale‌: Convert a color image to grayscale for easier processing.

[0054] Filtering: Use Gaussian filtering, mean filtering and other methods to remove image noise and smooth the image.

[0055] ‌Edge Detection‌: Apply edge detection operators such as Canny, Sobel, Prewitt, etc. to detect edges in the image.

[0056] 2. Contour Detection: On the basis of edge detection, the contours in the image are extracted through connected domain analysis, contour tracking and other methods.

[0057] Commonly used contour detection algorithms include Hough Transform for line and circle detection, as well as more complex shape matching algorithms.

[0058] 3. Contour processing: The extracted contours are further processed, such as contour smoothing, polygon approximation, contour simplification, etc., to improve recognition accuracy and efficiency.

[0059] ‌Contour Recognition and Classification‌: Identify and classify contours using methods such as template matching, machine learning (such as support vector machines, neural networks), or deep learning (such as convolutional neural networks).

[0060] Contour features (such as perimeter, area, shape descriptors, etc.) are often used as the basis for classification.

[0061] 4. Commonly used contour recognition algorithms ‌Canny Edge Detection‌: A multi-stage edge detection algorithm includes filtering, gradient calculation, non-maximum suppression and double threshold detection steps.

[0062] It can generate continuous, single-pixel-wide edges and is a commonly used edge detection method in contour recognition.

[0063] ‌Hough Transform‌: It is mainly used to detect geometric shapes such as straight lines and circles in images.

[0064] By finding the peak in the parameter space, shape detection and location are achieved.

[0065] ‌Shape Context‌: In this method, image recognition processing of each surgical instrument slice image can be performed in accordance with the above-mentioned image slicing method. The model can also perform batch image recognition on the collected images at one time. The model can perform image recognition on the image areas of each medical device contained in the image, identify the morphological features of several medical surgical instruments contained in the image at one time, and match and output the corresponding surgical instrument information.

[0066] In order to reduce the computational pressure of the model and improve recognition accuracy, the present invention preferentially adopts image slicing to perform batch image input (slices are sequentially input into the model for recognition processing) and recognition processing.

[0067] The morphological features are annotated with corresponding surgical instrument information, such as its attribute information, which may include the name of the current surgical instrument and the department in which it is used, etc.; the supply chain information may be the supply chain information of the surgical instrument, including the operation, warehousing, disinfection, transmission and other supply information of the surgical instrument; ownership information, that is, the department or management department to which the surgical instrument belongs, etc.; disinfection supply information, that is, the disinfection record information of the surgical instrument in and out of the warehouse, etc.; dynamic price information can be marked with prices to identify the prices of the surgical instruments currently in use, so as to facilitate the statistical billing of the prices of some special surgical instruments.

[0068] As an optional implementation scheme of the present application, optionally, S3, importing the surgical instrument image into the surgical instrument recognition model, identifying morphological features in the surgical instrument image by the surgical instrument recognition model, and outputting surgical instrument information corresponding to the morphological features, includes: Reading the image slices of each surgical instrument in sequence and inputting them into the surgical instrument recognition model; The morphological features in the image slices are identified by the surgical instrument recognition model, and the surgical instrument information marked on the morphological features is matched and output.

[0069] Each slice represents a surgical instrument.

[0070] Each slice is input into the model in turn, and the model recognizes the morphological features in the slice image and matches the surgical instrument information corresponding to the morphological features, so that the model can identify the type of surgical instrument in the current slice.

[0071] According to the above method, each slice image is subjected to such recognition processing, so that the system can recognize the surgical instrument information such as the name of each surgical instrument in the image, thereby achieving the purpose of rapid inventory.

[0072] Please understand and implement it in conjunction with the above model training and application steps.

[0073] As an optional implementation scheme of the present application, optionally, S4, counting the surgical instrument information and generating a corresponding surgical instrument inventory report, includes: Collecting statistics of surgical instrument information corresponding to the image slices of each surgical instrument; Writing the surgical instrument information into a preset inventory report to generate the surgical instrument inventory report for the surgical instruments currently located in the surgical instrument frame; The surgical instrument inventory report is saved and sent to the PDA terminal of the on-site nurse.

[0074] For the morphological features of each surgical instrument identified by the model and the surgical instrument information that matches it, the background can count the surgical instrument information output by the model that matches the corresponding morphological features, so as to collect and obtain the inventory information of each surgical instrument in the image collected this time, and write this information into the inventory report form or template preset in the background, and generate the inventory report of the surgical instruments counted this time in turn.

[0075] This report is an inventory report of each surgical instrument placed in the current surgical instrument frame. The background can store the report of this batch and send it to the handheld PDA terminal of the corresponding inventory personnel. The inventory personnel can log in to the background through the PDA terminal and view the corresponding inventory report.

[0076] For each batch of reports, the system can also record the sampling time and corresponding batch information, etc., and bind the corresponding sampling time, batch information and corresponding report to facilitate inventory management by the administrator.

[0077] Obviously, those skilled in the art should understand that the implementation of all or part of the processes in the above-mentioned embodiments can be completed by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the above-mentioned embodiments. Those skilled in the art can understand that the implementation of all or part of the processes in the above-mentioned embodiments can be completed by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the above-mentioned embodiments. Among them, the storage medium can be a disk, an optical disk, a read-only memory (ROM), a random access memory (RAM), a flash memory (Flash Memory), a hard disk (Hard Disk Drive, abbreviated as: HDD) or a solid-state drive (SSD), etc.; the storage medium can also include a combination of the above-mentioned types of memory. Example 2

[0078] like Figure 3 As shown, based on the implementation principle of Example 1, on the other hand, the present application proposes an intelligent surgical instrument inventory system for implementing the above-mentioned surgical instrument one-time inventory method, including: An image acquisition device, used to batch-acquire images of surgical instruments in the surgical instrument frame and upload them to the nurse station backend server; The nurse station backend server is used to import the surgical instrument image into the surgical instrument recognition model, identify the morphological features in the surgical instrument image through the surgical instrument recognition model, and output surgical instrument information corresponding to the morphological features; and count the surgical instrument information and generate a corresponding surgical instrument inventory report; PDA terminal, used to log in to the nurse station backend server, receive and view the surgical instrument inventory report; The image acquisition device and the PDA terminal are respectively connected to the backend server of the nurse station for communication.

[0079] Please understand and implement this system in conjunction with the corresponding interaction steps in Example 1, and this example will not be repeated.

[0080] The modules or steps of the present invention described above can be implemented by a general-purpose computing system, they can be concentrated on a single computing system, or distributed on a network composed of multiple computing systems, and optionally, they can be implemented by a program code executable by a computing system, so that they can be stored in a storage system and executed by the computing system, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. Thus, the present invention is not limited to any specific combination of hardware and software. Example 3

[0081] like Figure 4 As shown, further, in another aspect, the present application also proposes an electronic device, including: processor; a memory for storing processor-executable instructions; Wherein, the processor is configured to implement a one-time inventory method for surgical instruments as described in Example 1 when executing the executable instructions.

[0082] The electronic device of the embodiment of the present disclosure includes a processor and a memory for storing instructions executable by the processor, wherein the processor is configured to implement a one-time counting method for surgical instruments described in the above embodiment 1 when executing the executable instructions.

[0083] Here, it should be noted that the number of processors can be one or more. At the same time, the electronic device of the embodiment of the present disclosure may also include an input system and an output system. Among them, the processor, memory, input system and output system may be connected through a bus or in other ways, which are not specifically limited here.

[0084] The memory, as a computer-readable storage medium, can be used to store software programs, computer executable programs and various modules, such as the program or module corresponding to the one-time inventory method of surgical instruments in the embodiment of the present disclosure. The processor executes various functional applications and data processing of the electronic device by running the software programs or modules stored in the memory.

[0085] The input system can be used to receive input numbers or signals. The signal can be a key signal related to user settings and function control of the device / terminal / server. The output system can include display devices such as display screens.

[0086] The embodiments of the present disclosure have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or technical improvements in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

1. A one-time counting method for surgical instruments, characterized in that: The steps include: S1. Based on the surgical instrument image set, train and build a surgical instrument recognition model; S2, batch acquisition of surgical instrument images located in the surgical instrument frame; S3, importing the surgical instrument image into the surgical instrument recognition model, identifying morphological features in the surgical instrument image through the surgical instrument recognition model, and outputting surgical instrument information corresponding to the morphological features; S4. Count the surgical instrument information and generate a corresponding surgical instrument inventory report.

2. The one-time counting method for surgical instruments according to claim 1, characterized in that: S1. Based on the surgical instrument image set, train and build a surgical instrument recognition model, including: Collecting a number of surgical instrument images and forming the surgical instrument image set after image preprocessing; Performing feature engineering on the surgical instrument image set, and extracting morphological features of each surgical instrument image in the surgical instrument image set using CNN; The morphological features are marked as surgical instruments, where the surgical instrument information includes at least one of the following information: surgical instrument attribute information, supply chain information, ownership information, sterilization supply information or dynamic price information; After the labeling is completed, a feature set consisting of a number of labeled morphological features is obtained; Dividing the feature set into a training set and a validation set according to a preset ratio; Importing the training set into a preset CNN model, performing CNN model training, and generating the surgical instrument recognition model; The recognition performance of the surgical instrument recognition model is verified using the verification set: If the verification is successful, the surgical instrument recognition model is deployed on the nurse station backend server and put into use; Otherwise, repeat the training according to the above steps.

3. The one-time counting method for surgical instruments according to claim 2, characterized in that: S2. Batch acquisition of surgical instrument images located in the surgical instrument frame, including: Using an image acquisition device, photograph a number of surgical instruments laid out in a surgical instrument frame, batch-acquire surgical instrument images containing each surgical instrument, and upload them to a backend server at the nurse station; Using a contour recognition algorithm to identify the image contours of each surgical instrument in the surgical instrument image; Marking the image contours where each surgical instrument is located according to the identified image contours, and extracting image slices of each surgical instrument in the marked area; The image slices of each surgical instrument are sequentially saved in the database of the nurse station backend server.

4. The one-time counting method for surgical instruments according to claim 3, characterized in that: S3, importing the surgical instrument image into the surgical instrument recognition model, identifying the morphological features in the surgical instrument image through the surgical instrument recognition model, and outputting surgical instrument information corresponding to the morphological features, including: Reading the image slices of each surgical instrument in sequence and inputting them into the surgical instrument recognition model; The morphological features in the image slices are identified by the surgical instrument recognition model, and the surgical instrument information marked on the morphological features is matched and output.

5. The one-time counting method for surgical instruments according to claim 4, characterized in that: S4. Counting the surgical instrument information and generating a corresponding surgical instrument inventory report, including: Collecting statistics of surgical instrument information corresponding to the image slices of each surgical instrument; Writing the surgical instrument information into a preset inventory report to generate the surgical instrument inventory report for the surgical instruments currently located in the surgical instrument frame; The surgical instrument inventory report is saved and sent to the PDA terminal of the on-site nurse.

6. An intelligent surgical instrument inventory system, used to implement the one-time inventory method of surgical instruments according to any one of claims 1 to 5, characterized in that: include: An image acquisition device, used to batch-acquire images of surgical instruments in the surgical instrument frame and upload them to the nurse station backend server; The nurse station backend server is used to import the surgical instrument image into the surgical instrument recognition model, identify the morphological features in the surgical instrument image through the surgical instrument recognition model, and output surgical instrument information corresponding to the morphological features; and count the surgical instrument information and generate a corresponding surgical instrument inventory report; PDA terminal, used to log in to the nurse station backend server, receive and view the surgical instrument inventory report; The image acquisition device and the PDA terminal are respectively connected to the backend server of the nurse station for communication.

7. An electronic device, characterized in that include: processor; a memory for storing processor-executable instructions; Wherein, the processor is configured to implement a one-time inventory method for surgical instruments according to any one of claims 1-5 when executing the executable instructions.

Images