An automatic recognition system for surgical procedures based on video analysis

Abstract

The present invention provides a surgical procedure automatic recognition system based on video analysis, which belongs to the field of artificial intelligence and medical application technology. It includes a data acquisition module for acquiring surgical step accuracy data and surgical procedure recognition accuracy data; a surgical step accuracy index acquisition module for acquiring surgical step accuracy index; a surgical procedure recognition accuracy index acquisition module for acquiring surgical procedure recognition accuracy index; a data analysis module for acquiring surgical procedure automatic recognition system performance index; and a feedback module for feeding back the performance of the surgical procedure automatic recognition system. The present invention achieves the effect of accurately analyzing the results of the surgical procedure automatic recognition system by analyzing the proximity of the recognition results of the surgical procedure automatic recognition system to the simulated surgical operation process according to the performance index of the surgical procedure automatic recognition system, thereby solving the problem of inaccurate analysis of the results of the surgical procedure automatic recognition system in the prior art.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence and medical application technology, and in particular to an automatic surgical procedure recognition system based on video analysis. Background Art

[0002] With the continuous development of artificial intelligence in the medical field, the application of artificial intelligence covers multiple aspects such as preoperative diagnosis, clinical risk prediction, intraoperative decision support and postoperative monitoring. The automatic surgical process recognition system based on video analysis can identify the ongoing surgical tasks and remind possible complications. With the increase in surgical needs, the automatic surgical process recognition system based on video analysis plays a huge role in optimizing surgical processes and improving surgical results. Therefore, it is very important to check the accuracy of the recognition results of the automatic surgical process recognition system before using it. Analyzing the automatic surgical process recognition results is also of great significance to improving the accuracy of automatic surgical process recognition.

[0003] The existing automatic recognition system of surgical procedures extracts spatial features from videos through convolutional autoencoders, and then designs a convolutional temporal discriminant network to learn the sorting task of surgical videos based on the relatively fixed logical characteristics of surgical data, and obtains an unsupervised training network model based on convolutional autoencoders and temporal discriminant networks. The spatial features and temporal features of surgical videos are learned through unsupervised training of the network model, and the surgical video process can be determined based on the spatial features and temporal features. The loss function of the convolutional temporal discriminant network is adjusted to a softmax function, and semi-supervised learning is performed on the surgical video data. The surgical process is identified and the surgical video process stage identification results are output, thereby realizing the automatic recognition of the surgical process and analyzing the identification results. However, in the prior art, due to the diversity and complexity of the surgical process, and the analysis of the recognition results in the prior art is not comprehensive enough, resulting in low accuracy of automatic recognition of the surgical process, and inaccurate analysis of the results of automatic recognition of the surgical process. Summary of the invention

[0004] The embodiment of the present invention solves the problem of inaccurate analysis of the results of automatic recognition of surgical procedures in the prior art by providing an automatic recognition system for surgical procedures based on video analysis, and realizes accurate analysis of the results of automatic recognition of surgical procedures. The technical solution is as follows:

[0005] A surgical procedure automatic recognition system based on video analysis, comprising: a data acquisition module, a surgical step accuracy index acquisition module, a surgical procedure recognition accuracy index acquisition module, a data analysis module and a feedback module; wherein the data acquisition module is used to acquire surgical step accuracy data and surgical procedure recognition accuracy data; the surgical step accuracy index acquisition module is used to acquire surgical step accuracy indicators based on surgical step accuracy data; the surgical procedure recognition accuracy index acquisition module is used to acquire surgical procedure recognition accuracy indicators based on surgical procedure recognition accuracy data; the data analysis module is used to acquire surgical procedure automatic recognition system performance indicators based on surgical step accuracy indicators and surgical procedure recognition accuracy indicators; the feedback module is used to provide feedback on the performance of the surgical procedure automatic recognition system based on the surgical procedure automatic recognition system performance indicators.

[0006] Optionally, the surgical step accuracy data includes surgical step recognition accuracy, surgical step recognition recall rate and surgical step recognition F1 score; the surgical procedure recognition accuracy data includes surgical procedure recognition accuracy standard data and surgical procedure recognition accuracy actual data.

[0007] Optionally, the standard data for surgical procedure recognition accuracy include standard data for surgical incision length, standard data for surgical time, standard data for surgical bleeding volume, and the number of times each instrument is used in the simulated surgical process; the actual data for surgical procedure recognition accuracy include actual data for surgical incision length, actual data for surgical time, actual data for surgical bleeding volume, and the number of times each instrument is used identified by the automatic surgical procedure recognition system based on video analysis.

[0008] Optionally, the surgical step recognition accuracy is the ratio of the number of process steps correctly recognized by the surgical process automatic recognition system to the total number of steps recognized by the surgical process automatic recognition system; the surgical step recognition recall rate is the ratio of the number of process steps correctly recognized by the surgical process automatic recognition system to the total number of steps in the simulated surgical process; the surgical step recognition F1 score is the harmonic mean of the surgical step recognition accuracy and the surgical step recognition recall rate.

[0009] Optionally, the specific analysis process of the surgical step accuracy index is: obtain the surgical step accuracy index, obtain a first threshold value of the surgical step accuracy index and a second threshold value of the surgical step accuracy index, compare the surgical step accuracy index with the first threshold value of the surgical step accuracy index and the second threshold value of the surgical step accuracy index, when the surgical step accuracy index is less than the first threshold value of the surgical step accuracy index and greater than the second threshold value of the surgical step accuracy index, it indicates that the surgical step accuracy index is in a normal range, that is, the surgical step accuracy index is in a passing range; when in other situations, it indicates that the surgical step accuracy index is in an abnormal range, that is, the surgical step accuracy index is in a failing range.

[0010] Optionally, the specific method for obtaining the surgical process identification accuracy index is: obtaining the types of instruments used in the surgical process and numbering the types of instruments used in the surgical process; obtaining preset surgical incision length standard data, surgical time standard data, surgical bleeding volume standard data and the weight ratio of the number of times each instrument is used in the simulated surgical process in the surgical process identification accuracy index from the database; constructing a surgical process identification accuracy index calculation formula; the specific surgical process identification accuracy index calculation formula is:

[0011]

[0012] In the formula, θ represents the surgical process recognition accuracy index, a represents the actual data of surgical incision length, a1 represents the standard data of surgical incision length, b represents the actual data of surgical time, b1 represents the standard data of surgical time, c represents the actual data of surgical bleeding, c1 represents the standard data of surgical bleeding, and e represents the actual data of surgical bleeding. i It represents the number of times the i-th device identified by the system is used, i represents the number of the device type, i = 1, 2, ..., n, n represents the total number of device types, e i ′ It is represented by the number of times the i-th instrument is used in the simulated surgical process, α is represented by the accuracy index of the surgical steps, d1 is represented by the weight ratio of the actual data of the surgical incision length in the surgical process recognition accuracy index, d2 is represented by the weight ratio of the actual data of the surgical time in the surgical process recognition accuracy index, d3 is represented by the weight ratio of the actual data of the surgical bleeding volume in the surgical process recognition accuracy index, d4 is represented by the weight ratio of the number of times each instrument is used identified by the system in the surgical process recognition accuracy index, and d5 is represented by the weight ratio of the surgical step accuracy index in the surgical process recognition accuracy index.

[0013] Optionally, the specific analysis process of the surgical procedure identification accuracy index is: obtaining the surgical procedure identification accuracy index, obtaining a first threshold value of the surgical procedure identification accuracy index and a second threshold value of the surgical procedure identification accuracy index, comparing the surgical procedure identification accuracy index with the first threshold value of the surgical procedure identification accuracy index and the second threshold value of the surgical procedure identification accuracy index; when the surgical procedure identification accuracy index is less than the first threshold value of the surgical procedure identification accuracy index and greater than the second threshold value of the surgical procedure identification accuracy index, it indicates that the surgical procedure identification accuracy index is in a normal range, that is, the surgical procedure identification accuracy index is in a passing range; when in other situations, it indicates that the surgical procedure identification accuracy index is in an abnormal range, that is, the surgical procedure identification accuracy index is in a failing range.

[0014] Optionally, the specific method for obtaining the performance indicators of the surgical procedure automatic identification system is:

[0015]

[0016] Wherein, γ represents the performance index of the automatic recognition system for surgical procedures, α represents the accuracy index of surgical steps, θ represents the accuracy index of surgical procedure recognition, w1 represents the weight ratio of the accuracy index of surgical steps in the performance index of the automatic recognition system for surgical procedures, w2 represents the weight ratio of the accuracy index of surgical procedure recognition in the performance index of the automatic recognition system for surgical procedures, and e is a natural constant.

[0017] Optionally, the specific analysis process of the performance indicator of the automatic surgical procedure identification system is: obtaining the performance indicator of the automatic surgical procedure identification system, obtaining the first threshold value of the performance indicator of the automatic surgical procedure identification system and the second threshold value of the performance indicator of the automatic surgical procedure identification system, comparing the performance indicator of the automatic surgical procedure identification system with the first threshold value of the performance indicator of the automatic surgical procedure identification system and the second threshold value of the performance indicator of the automatic surgical procedure identification system, when the performance indicator of the automatic surgical procedure identification system is less than the first threshold value of the performance indicator of the automatic surgical procedure identification system and greater than the second threshold value of the performance indicator of the automatic surgical procedure identification system, it indicates that the performance indicator of the automatic surgical procedure identification system is in a normal range, that is, the performance indicator of the automatic surgical procedure identification system is in a passing range; when in other situations, it indicates that the performance indicator of the automatic surgical procedure identification system is in an abnormal range, that is, the performance indicator of the automatic surgical procedure identification system is in a failing range.

[0018] Optionally, the feedback module provides feedback on the performance of the automatic surgical procedure recognition system based on video analysis according to the performance indicators of the automatic surgical procedure recognition system. When the performance indicators of the automatic surgical procedure recognition system are in an unsatisfactory range, the recognition results of the automatic surgical procedure recognition system fed back deviate greatly from the simulated surgical operation process; when the performance indicators of the automatic surgical procedure recognition system are in a passing range, the recognition results of the automatic surgical procedure recognition system fed back are more consistent with the simulated surgical operation process.

[0019] Compared with the prior art, the above technical solution has at least the following beneficial effects:

[0020] 1. By acquiring the surgical step accuracy data and the surgical process recognition accuracy data, and analyzing to obtain the surgical step accuracy index and the surgical process recognition accuracy index, and then obtaining the surgical process automatic recognition system performance index based on the surgical step accuracy index and the surgical process recognition accuracy index, the degree of proximity between the recognition results of the surgical process automatic recognition system and the simulated surgical operation process is analyzed based on the surgical process automatic recognition system performance index, thereby achieving accurate analysis of the results of the surgical process automatic recognition, and effectively solving the problem of inaccurate analysis of the results of the surgical process automatic recognition in the prior art.

[0021] 2. By obtaining the surgical step recognition accuracy, surgical step recognition recall rate and surgical step recognition F1 score, the surgical step accuracy index is obtained, and then the surgical step accuracy is analyzed according to the surgical step accuracy index, thereby achieving accurate analysis of the surgical step accuracy.

[0022] 3. By acquiring standard data on surgical incision length, standard data on surgical time, standard data on surgical bleeding volume, the number of times each instrument is used in the simulated surgical process, actual data on surgical incision length, actual data on surgical time, actual data on surgical bleeding volume and the number of times each instrument is used as identified by the system, and analyzing them to obtain surgical process recognition accuracy indicators, the surgical process recognition accuracy indicators are analyzed according to the surgical process recognition accuracy indicators, thereby achieving accurate analysis of surgical process recognition accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0024] Figure 1 A diagram of an automatic surgical procedure recognition system based on video analysis provided by an embodiment of the present invention;

[0025] Figure 2 This is a performance indicator curve chart of the automatic surgical procedure identification system provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0026] In order to make the purpose, technical solution and advantages of the embodiment of the present invention clearer, the technical solution of the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings of the embodiment of the present invention. Obviously, the described embodiment is a part of the embodiment of the present invention, not all of the embodiments. Based on the described embodiment of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0027] Unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The words "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. Similarly, words such as "one", "one" or "the" do not indicate a quantitative limitation, but indicate the existence of at least one. Words such as "include" or "comprise" mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. Words such as "connect" or "connected" are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect.

[0028] It should be noted that the terms "upper", "lower", "left", "right", "front", "back", etc. used in the present invention are only used to indicate relative positional relationships. When the absolute position of the object being described changes, the relative positional relationship may also change accordingly.

[0029] The embodiment of the present invention solves the problem of inaccurate analysis of the results of automatic recognition of surgical procedures in the prior art by providing an automatic recognition system for surgical procedures based on video analysis. By acquiring surgical step accuracy data and surgical procedure recognition accuracy data, surgical step accuracy indicators and surgical procedure recognition accuracy indicators are obtained, and then the performance indicators of the automatic recognition system for surgical procedures are obtained based on the surgical step accuracy indicators and surgical procedure recognition accuracy indicators. The degree of proximity between the recognition results of the automatic recognition system for surgical procedures and the simulated surgical operation process is fed back based on the performance indicators of the automatic recognition system for surgical procedures, thereby achieving accurate analysis of the results of automatic recognition of surgical procedures, and effectively solving the problem of inaccurate analysis of the results of automatic recognition of surgical procedures in the prior art.

[0030] The technical solution in the embodiment of the present invention is to solve the problem of inaccurate analysis of the results of automatic recognition of surgical procedures. The overall idea is as follows:

[0031] The surgical step recognition accuracy, surgical step recognition recall and surgical step recognition F1 score are obtained to obtain the surgical step accuracy index; the surgical process recognition accuracy index is obtained by obtaining standard data on surgical incision length, standard data on surgical time, standard data on surgical bleeding volume, the number of times each instrument is used in the simulated surgical process, actual data on surgical incision length, actual data on surgical time, actual data on surgical bleeding volume and the number of times each instrument is used identified by the system; the surgical step accuracy index and the surgical process recognition accuracy index are combined to obtain the performance index of the automatic surgical process recognition system; the performance index of the automatic surgical process recognition system is analyzed to see how close the recognition results of the automatic surgical process recognition system are to the simulated surgical operation process, thereby achieving the effect of accurately analyzing the results of automatic surgical process recognition.

[0032] like Figure 1 As shown, a diagram of a surgical procedure automatic recognition system based on video analysis provided by an embodiment of the present invention includes: a data acquisition module, a surgical step accuracy index acquisition module, a surgical procedure recognition accuracy index acquisition module, a data analysis module and a feedback module; wherein the data acquisition module is used to acquire surgical step accuracy data and surgical procedure recognition accuracy data; the surgical step accuracy index acquisition module is used to acquire surgical step accuracy indicators based on surgical step accuracy data; the surgical procedure recognition accuracy index acquisition module is used to acquire surgical procedure recognition accuracy indicators based on surgical procedure recognition accuracy data; the data analysis module is used to acquire surgical procedure automatic recognition system performance indicators based on surgical step accuracy indicators and surgical procedure recognition accuracy indicators; the feedback module is used to provide feedback on the performance of the surgical procedure automatic recognition system based on the surgical procedure automatic recognition system performance indicators.

[0033] In this embodiment, by comprehensively evaluating the accuracy of surgical steps and surgical process recognition, the recognition results of the surgical process automatic recognition system are analyzed, which is conducive to more accurate analysis of the recognition results of the surgical process automatic recognition system.

[0034] In this embodiment, the surgical step accuracy data and the surgical process recognition accuracy data need to be obtained in the simulated surgery, the incision length and the bleeding volume of the simulated surgery need to be set, the simulated surgery is performed on the simulation model, and the simulated surgery video is obtained. The set simulated surgery incision length and the bleeding volume of the simulated surgery are the surgical incision length standard data and the surgical bleeding volume standard data. It should be noted that the videos mentioned in this embodiment are all simulated surgery videos.

[0035] Optionally, the surgical step accuracy data includes surgical step recognition accuracy, surgical step recognition recall rate and surgical step recognition F1 score; the surgical procedure recognition accuracy data includes surgical procedure recognition accuracy standard data and surgical procedure recognition accuracy actual data.

[0036] In this embodiment, the surgical step recognition accuracy is obtained by dividing the number of process steps correctly recognized by the surgical process automatic recognition system by the total number of steps recognized by the surgical process automatic recognition system. The surgical step recognition recall rate is obtained by dividing the number of process steps correctly recognized by the surgical process automatic recognition system by the total number of steps in the simulated surgical process. The surgical step recognition F1 score is obtained by calculating the sum of the inverse of the surgical step recognition accuracy and the inverse of the surgical step recognition recall rate, and then dividing the sum by 2 to obtain the surgical step recognition F1 score.

[0037] In the present embodiment, it can be known from the above that the surgical step recognition accuracy, surgical step recognition recall rate and surgical step recognition F1 score are obtained by calculating the number of process steps correctly recognized by the surgical process automatic recognition system, the total number of steps recognized by the surgical process automatic recognition system and the total number of steps in the simulated surgical process. Therefore, in the present embodiment, it is necessary to obtain the number of process steps correctly recognized by the surgical process automatic recognition system, the total number of steps recognized by the surgical process automatic recognition system and the total number of steps in the simulated surgical process. The number of process steps correctly recognized by the surgical process automatic recognition system is obtained by counting and recording the number of process steps correctly recognized in the surgical video of the medical staff, the total number of steps recognized by the surgical process automatic recognition system is obtained by recording the total number of steps recognized by the surgical process automatic recognition system in the surgical video, and the total number of steps in the simulated surgical process is obtained by recording the number of steps actually operated by the medical staff in the surgical process.

[0038] Optionally, the standard data for surgical procedure recognition accuracy include standard data for surgical incision length, standard data for surgical time, standard data for surgical bleeding volume, and the number of times each instrument is used in the simulated surgical process; the actual data for surgical procedure recognition accuracy include actual data for surgical incision length, actual data for surgical time, actual data for surgical bleeding volume, and the number of times each instrument is used identified by the automatic surgical procedure recognition system based on video analysis.

[0039] In this embodiment, the standard data of surgical incision length is obtained by directly measuring the surgical incision during the simulated operation, the standard data of surgical bleeding volume is obtained by collecting the bleeding volume during the simulated operation, the standard data of operation time is obtained by the difference between the start time and the end time of the simulated operation, the number of times each instrument is used in the simulated operation process is recorded by medical personnel during the operation, the actual data of operation time is obtained by the difference between the start time and the end time of the operation identified by the automatic recognition system of the operation process based on video analysis, the actual data of surgical incision length can be measured by the automatic recognition system of the operation process through image processing technology, the actual data of surgical bleeding volume is obtained by distinguishing blood and estimating the bleeding volume through the image analysis technology of the automatic recognition system of the operation process, and the number of times each instrument is used identified by the system is obtained by statistically recording the number of times each instrument is identified in the surgical process video by the automatic recognition system of the operation process.

[0040] Optionally, the surgical step recognition accuracy is the ratio of the number of process steps correctly recognized by the surgical process automatic recognition system to the total number of steps recognized by the surgical process automatic recognition system; the surgical step recognition recall is the ratio of the number of process steps correctly recognized by the surgical process automatic recognition system to the total number of steps in the simulated surgical process; the surgical step recognition F1 score is the harmonic mean of the surgical step recognition accuracy and the surgical step recognition recall.

[0041] In this embodiment, the surgical step recognition accuracy measures the ability of the surgical process automatic recognition system to correctly identify surgical steps, the surgical step recognition recall measures the system's ability to recognize all actual surgical steps, and the surgical step recognition F1 score represents the overall situation of surgical step recognition.

[0042] Optionally, the specific analysis process of the surgical step accuracy index is: obtain the surgical step accuracy index, obtain a first threshold value of the surgical step accuracy index and a second threshold value of the surgical step accuracy index, compare the surgical step accuracy index with the first threshold value of the surgical step accuracy index and the second threshold value of the surgical step accuracy index, when the surgical step accuracy index is less than the first threshold value of the surgical step accuracy index and greater than the second threshold value of the surgical step accuracy index, it indicates that the surgical step accuracy index is in a normal range, that is, the surgical step accuracy index is in a passing range; when in other situations, it indicates that the surgical step accuracy index is in an abnormal range, that is, the surgical step accuracy index is in a failing range.

[0043] In this embodiment, the specific method for obtaining the surgical step accuracy index is:

[0044] Obtain from the database the number of process steps correctly identified by the preset surgical process automatic recognition system, the total number of simulated surgical process steps, and the weight ratio of the total number of steps identified by the surgical process automatic recognition system in the surgical step accuracy index; construct a surgical step accuracy index calculation formula; the specific surgical step accuracy index calculation formula is:

[0045]

[0046] Wherein, α is the surgical step accuracy index, s is the number of process steps correctly identified by the surgical process automatic recognition system, s1 is the total number of steps identified by the surgical process automatic recognition system, s2 is the total number of simulated surgical process steps, h1 is the weight ratio of the number of process steps correctly identified by the surgical process automatic recognition system in the surgical step accuracy index, h2 is the weight ratio of the total number of simulated surgical process steps in the surgical step accuracy index, h3 is the weight ratio of the total number of steps identified by the surgical process automatic recognition system in the surgical step accuracy index, and e is a natural constant.

[0047] In this embodiment, s represents the number of correct process steps recognized by the automatic surgical process recognition system based on video analysis. In this embodiment, the number of process steps correctly recognized by the automatic surgical process recognition system is the number of steps recognized by the automatic surgical process recognition system that match the names of actual operation steps. s1 represents the total number of steps recognized by the automatic surgical process recognition system based on video analysis, and s2 represents the total number of steps in the actual surgical operation process.

[0048] In this embodiment, h1 is the influence weight ratio of the number of process steps correctly identified by the preset automatic surgical process recognition system obtained from the preset database, which is used to represent the numerical value of the influence of the number of process steps correctly identified by the automatic surgical process recognition system on the surgical step accuracy index. When used, the influence weight ratio of the number of process steps correctly identified by the automatic surgical process recognition system corresponding to the number of process steps correctly identified by the automatic surgical process recognition system can be directly obtained from the preset database. The corresponding relationship can be a pre-set mapping relationship. For example, the number of process steps correctly identified by the automatic surgical process recognition system and the weight corresponding to the number of process steps correctly identified by the automatic surgical process recognition system in the preset database form a mapping set. After the real-time number of process steps correctly identified by the automatic surgical process recognition system is input into the mapping set, the corresponding weight will be obtained. In this example, its value range is [0, 1]. h2 is the influence weight ratio of the preset total number of steps in the simulated surgical process obtained from the preset database, which is used to represent the numerical value of the influence of the total number of steps in the simulated surgical process on the surgical step accuracy index. When used, the influence weight ratio of the total number of steps in the simulated surgical process corresponding to the total number of steps in the simulated surgical process can be directly obtained from the preset database. The corresponding relationship can be a preset mapping relationship. For example, the total number of steps in the simulated surgical process and the weight corresponding to the total number of steps in the simulated surgical process in the preset database form a mapping set. After the real-time total number of steps in the simulated surgical process is input into the mapping set, the corresponding weight will be obtained. In this example, its value range is [0, 1]. h3 is the influence weight ratio of the total number of steps identified by the preset automatic recognition system of the surgical procedure obtained from the preset database, which is used to represent the numerical value of the influence degree of the total number of steps identified by the automatic recognition system of the surgical procedure on the surgical step accuracy index. When used, the influence weight ratio of the total number of steps identified by the automatic recognition system of the surgical procedure corresponding to the total number of steps identified by the automatic recognition system of the surgical procedure can be directly obtained from the preset database. The corresponding relationship can be a preset mapping relationship. For example, the total number of steps identified by the automatic recognition system of the surgical procedure and the weight corresponding to the total number of steps identified by the automatic recognition system of the surgical procedure in the preset database form a mapping set. After the total number of steps identified by the automatic recognition system of the surgical procedure in real time is input into the mapping set, the corresponding weight will be obtained. In this example, its value range is [0, 1].

[0049] In this embodiment, the surgical step accuracy index is calculated by the above formula for the number of process steps correctly identified by the surgical process automatic recognition system, the total number of steps identified by the surgical process automatic recognition system, and the total number of steps in the simulated surgical process. The closer the number of process steps correctly identified by the surgical process automatic recognition system and the total number of steps identified by the surgical process automatic recognition system are to the total number of steps in the simulated surgical process, the better the surgical step accuracy index is. The above formula uses square root processing to smooth the influence of outliers, so that the obtained surgical step accuracy index is more effective and reliable.

[0050] In this embodiment, historical data of parameters related to the surgical step accuracy index are obtained from the database, the historical data of the surgical step accuracy index is obtained according to the above formula, the historical data of the surgical step accuracy index is averaged, the average is multiplied by 1.1 to obtain the first threshold of the surgical step accuracy index, and the first threshold of the surgical step accuracy index is α1. The average is multiplied by 0.9 to obtain the second threshold of the surgical step accuracy index, and the second threshold of the surgical step accuracy index is α2.

[0051] In this embodiment, α1=1.05, α2=0.95. When the surgical step accuracy index is less than the first threshold value of the surgical step accuracy index and greater than the second threshold value of the surgical step accuracy index, that is, α2<α<α1, it indicates that the surgical step accuracy index is in the normal range, that is, the surgical step accuracy is in the passing range; when in other situations, that is, α≥α1 or α≤α2, it indicates that the surgical step accuracy index is in the abnormal range, that is, the surgical step accuracy is in the failing range.

[0052] Optionally, the specific method for obtaining the surgical process recognition accuracy index is: obtaining the types of instruments used in the surgical process and numbering the types of instruments used in the surgical process; obtaining preset surgical incision length standard data, surgical time standard data, surgical bleeding volume standard data and the weight ratio of the number of times each instrument is used in the simulated surgical process in the surgical process recognition accuracy index from the database; constructing a surgical process recognition accuracy index calculation formula; the specific surgical process recognition accuracy index calculation formula is:

[0053]

[0054] In the formula, θ represents the surgical process recognition accuracy index, a represents the actual data of surgical incision length, a1 represents the standard data of surgical incision length, b represents the actual data of surgical time, b1 represents the standard data of surgical time, c represents the actual data of surgical bleeding, c1 represents the standard data of surgical bleeding, and e represents the actual data of surgical bleeding. iIt represents the number of times the i-th device identified by the system is used, i represents the number of the device type, i = 1, 2, ..., n, n represents the total number of device types, e i ′ It is represented by the number of times the i-th instrument is used in the simulated surgical process, α is represented by the accuracy index of the surgical steps, d1 is represented by the weight ratio of the actual data of the surgical incision length in the surgical process recognition accuracy index, d2 is represented by the weight ratio of the actual data of the surgical time in the surgical process recognition accuracy index, d3 is represented by the weight ratio of the actual data of the surgical bleeding volume in the surgical process recognition accuracy index, d4 is represented by the weight ratio of the number of times each instrument is used identified by the system in the surgical process recognition accuracy index, and d5 is represented by the weight ratio of the surgical step accuracy index in the surgical process recognition accuracy index.

[0055] In this embodiment, a represents the length of the surgical incision during the simulated surgical operation, a1 represents the length of the incision of the simulated surgery preset by the medical team, b represents the difference between the start time and the end time of the surgery identified by the automatic surgical process recognition system, b1 represents the difference between the start time and the end time of the simulated surgery, c represents the amount of bleeding identified by the automatic surgical process recognition system, c1 represents the amount of bleeding of the simulated surgery preset by the medical team, and e i represents the number of times each instrument is used during the simulated surgical operation, e i ′ Represents the number of times each instrument is identified by the surgical procedure automatic recognition system.

[0056] In this embodiment, d1 is the influence weight ratio of the preset actual data of the surgical incision length obtained from the preset database, which is used to represent the numerical value of the influence degree of the actual data of the surgical incision length on the surgical process recognition accuracy index; d2 is the influence weight ratio of the preset actual data of the surgical time obtained from the preset database, which is used to represent the numerical value of the influence degree of the actual data of the surgical time on the surgical process recognition accuracy index; d3 is the influence weight ratio of the preset actual data of the surgical bleeding volume obtained from the preset database, which is used to represent the numerical value of the influence degree of the actual data of the surgical bleeding volume on the surgical process recognition accuracy index; d4 is the influence weight ratio of the preset number of uses of each instrument identified by the system obtained from the preset database, which is used to represent the numerical value of the influence degree of the number of uses of each instrument identified by the system on the surgical process recognition accuracy index; d5 is the influence weight ratio of the preset surgical step accuracy index obtained from the preset database, which is used to represent the numerical value of the influence degree of the surgical step accuracy index on the surgical process recognition accuracy index. When in use, the weight ratio of the actual data of surgical incision length corresponding to the actual data of surgical incision length, the weight ratio of the actual data of surgical time corresponding to the actual data of surgical time, the weight ratio of the actual data of surgical bleeding corresponding to the actual data of surgical bleeding, the weight ratio of the number of times each instrument is used identified by the system and the weight ratio of the number of times the surgical step accuracy index is used corresponding to the surgical step accuracy index can be directly obtained from the preset database. The corresponding relationship can be a preset mapping relationship, for example, the actual data of surgical incision length, the actual data of surgical time, the actual data of surgical bleeding, the number of times each instrument is used identified by the system and the weight ratio of the surgical step accuracy index A mapping set is formed between the accuracy index and the actual data of surgical incision length, actual data of surgical time, actual data of surgical bleeding volume, the number of times each instrument is used identified by the system, and the weights corresponding to the accuracy index of surgical steps in a preset database. After the real-time actual data of surgical incision length, actual data of surgical time, actual data of surgical bleeding volume, the number of times each instrument is used identified by the system, and the weights of the accuracy index of surgical steps are input into the mapping set, the corresponding actual data of surgical incision length, actual data of surgical time, actual data of surgical bleeding volume, the number of times each instrument is used identified by the system, and the weights of the accuracy index of surgical steps will be obtained. In this example, the value ranges of d1, d2, d3, d4, and d5 are all [0, 1].

[0057] In this embodiment, various factors including actual data of surgical incision length, actual data of surgical time, actual data of surgical bleeding volume, number of times each instrument is used identified by the system and index of surgical step accuracy are comprehensively considered, and the influence of various factors is taken into account. The closer the actual data of surgical incision length, actual data of surgical time, actual data of surgical bleeding volume and number of times each instrument is used identified by the system are to standard data of surgical incision length, standard data of surgical time, standard data of surgical bleeding volume and number of times each instrument is used during the simulated surgical operation and the closer the surgical step accuracy index is to 1, the closer the surgical process recognition accuracy index is to 1. The above formula uses logarithmic operation to reduce the influence of outliers, so that the obtained surgical process recognition accuracy index is more accurate and reliable.

[0058] Optionally, the specific analysis process of the surgical procedure recognition accuracy index is: obtaining the surgical procedure recognition accuracy index, obtaining the first threshold value of the surgical procedure recognition accuracy index and the second threshold value of the surgical procedure recognition accuracy index, comparing the surgical procedure recognition accuracy index with the first threshold value of the surgical procedure recognition accuracy index and the second threshold value of the surgical procedure recognition accuracy index; when the surgical procedure recognition accuracy index is less than the first threshold value of the surgical procedure recognition accuracy index and greater than the second threshold value of the surgical procedure recognition accuracy index, it indicates that the surgical procedure recognition accuracy index is in a normal range, that is, the surgical procedure recognition accuracy index is in a passing range; when in other situations, it indicates that the surgical procedure recognition accuracy index is in an abnormal range, that is, the surgical procedure recognition accuracy index is in a failing range.

[0059] In this embodiment, historical data of parameters related to the surgical process recognition accuracy index are obtained from the database, the historical data of the surgical process recognition accuracy index is obtained according to the above formula, the historical data of the surgical process recognition accuracy index is mean calculated, the mean is multiplied by 1.1 to obtain the first threshold of the surgical process recognition accuracy index, the first threshold of the surgical process recognition accuracy index is θ1, the mean is multiplied by 0.9 to obtain the second threshold of the surgical process recognition accuracy index, the second threshold of the surgical process recognition accuracy index is θ2.

[0060] In this embodiment, θ1=1.04, θ2=0.95. When the surgical procedure recognition accuracy index is less than the first threshold value of the surgical procedure recognition accuracy index and greater than the second threshold value of the surgical procedure recognition accuracy index, that is, θ2<θ<θ1, it indicates that the surgical procedure recognition accuracy index is in the normal range, that is, the surgical procedure recognition accuracy is in the passing range; when in other situations, that is, θ≥θ1 or θ≤θ2, it indicates that the surgical procedure recognition accuracy index is in the abnormal range, that is, the surgical procedure recognition accuracy is in the failing range.

[0061] Optionally, a specific method for obtaining the performance indicators of the surgical procedure automatic identification system is:

[0062]

[0063] Wherein, γ represents the performance index of the automatic recognition system for surgical procedures, α represents the accuracy index of surgical steps, θ represents the accuracy index of surgical procedure recognition, w1 represents the weight ratio of the accuracy index of surgical steps in the performance index of the automatic recognition system for surgical procedures, w2 represents the weight ratio of the accuracy index of surgical procedure recognition in the performance index of the automatic recognition system for surgical procedures, and e is a natural constant.

[0064] In this embodiment, w1 is the influence weight ratio of the preset surgical step accuracy index obtained from the preset database, which is used to indicate the numerical value of the influence degree of the surgical step accuracy index on the surgical step accuracy index. When used, the influence weight ratio of the surgical step accuracy index corresponding to the surgical step accuracy index can be directly obtained from the preset database. The corresponding relationship can be a preset mapping relationship. For example, the surgical step accuracy index and the weight corresponding to the surgical step accuracy index in the preset database form a mapping set. After the real-time surgical step accuracy index is input into the mapping set, the corresponding weight will be obtained. In this example, the value range is [0, 1]. w2 is the influence weight ratio of the preset surgical process recognition accuracy index obtained from the preset database, which is used to indicate the numerical value of the influence degree of the surgical process recognition accuracy index on the surgical step accuracy index. When used, the influence weight ratio of the surgical process recognition accuracy index corresponding to the surgical process recognition accuracy index can be directly obtained from the preset database. The corresponding relationship can be a preset mapping relationship. For example, the surgical process recognition accuracy index and the weight corresponding to the surgical process recognition accuracy index in the preset database form a mapping set. After the real-time surgical process recognition accuracy index is input into the mapping set, the corresponding weight will be obtained. In this example, the value range is [0, 1].

[0065] In this embodiment, the surgical step accuracy index and the surgical process accuracy index are integrated to more accurately analyze the recognition accuracy of the surgical process automatic recognition system. The closer the surgical step accuracy index and the surgical process recognition accuracy index are to 1, the closer the surgical process automatic recognition system performance index is to 1. The closer the surgical process automatic recognition system performance index is to 1, the closer the recognition result of the surgical process automatic recognition system is to the simulated surgical operation process. The above formula uses a hyperbolic tangent function for nonlinear transformation, which increases the stability of the relevant data, smoothes the influence of outliers, and makes the obtained surgical process automatic recognition system performance index more accurate and effective.

[0066] like Figure 2As shown, it is a performance index curve diagram of the automatic surgical process identification system provided by an embodiment of the present invention, with the surgical step accuracy index as the independent variable and the surgical process automatic identification system performance index as the dependent variable. Through the increase and decrease analysis of curves 1, 2 and 3, it can be known that when the surgical step accuracy index is larger, the performance index of the automatic surgical process identification system is larger; when the surgical step accuracy index is smaller, the performance index of the automatic surgical process identification system is smaller. The statistical table of relevant data of the performance index of the automatic surgical process identification system is shown in Table 1:

[0067] Table 1 Statistics of performance indicators of the automatic recognition system for surgical procedures

[0068]

[0069] It can be seen from Table 1 that for curve 1, curve 2 and curve 3, when the surgical step accuracy index of curve 1, curve 2 and curve 3 is 1.00, when the surgical process recognition accuracy index of curve 1, curve 2 and curve 3 is 0.80, 1.00 and 1.20 respectively, and other parameters are consistent, the performance indexes of the surgical process automatic recognition system of curve 1, curve 2 and curve 3 are 0.95, 1.00 and 1.05 respectively. It can be seen that when the surgical step accuracy index is consistent with other parameters, when the surgical process recognition accuracy index is larger, the performance index of the surgical process automatic recognition system is larger.

[0070] Optionally, the specific analysis process of the performance indicator of the automatic surgical procedure identification system is: obtain the performance indicator of the automatic surgical procedure identification system, obtain the first threshold value of the performance indicator of the automatic surgical procedure identification system and the second threshold value of the performance indicator of the automatic surgical procedure identification system, compare the performance indicator of the automatic surgical procedure identification system with the first threshold value of the performance indicator of the automatic surgical procedure identification system and the second threshold value of the performance indicator of the automatic surgical procedure identification system, when the performance indicator of the automatic surgical procedure identification system is less than the first threshold value of the performance indicator of the automatic surgical procedure identification system and greater than the second threshold value of the performance indicator of the automatic surgical procedure identification system, it indicates that the performance indicator of the automatic surgical procedure identification system is in a normal range, that is, the performance indicator of the automatic surgical procedure identification system is in a passing range; when in other situations, it indicates that the performance indicator of the automatic surgical procedure identification system is in an abnormal range, that is, the performance indicator of the automatic surgical procedure identification system is in a failing range.

[0071] In this embodiment, historical data of parameters related to the performance indicators of the automatic surgical procedure identification system are obtained from the database, and the historical data of the performance indicators of the automatic surgical procedure identification system are obtained according to the above formula. The mean of the historical data of the performance indicators of the automatic surgical procedure identification system is calculated, and the mean is multiplied by 1.1 to obtain the first threshold value of the performance indicators of the automatic surgical procedure identification system, and the first threshold value of the performance indicators of the automatic surgical procedure identification system is. The mean is multiplied by 0.9 to obtain the second threshold value of the performance indicators of the automatic surgical procedure identification system, and the second threshold value of the performance indicators of the automatic surgical procedure identification system is.

[0072] In this embodiment, historical data of parameters related to the performance indicators of the automatic surgical procedure recognition system are obtained from the database, and the historical data of the performance indicators of the automatic surgical procedure recognition system are obtained according to the above formula. The historical data of the performance indicators of the automatic surgical procedure recognition system are mean calculated, and the mean is multiplied by 1.1 to obtain the first threshold value of the performance indicators of the automatic surgical procedure recognition system, and the first threshold value of the performance indicators of the automatic surgical procedure recognition system is γ1. The mean is multiplied by 0.9 to obtain the second threshold value of the performance indicators of the automatic surgical procedure recognition system, and the second threshold value of the performance indicators of the automatic surgical procedure recognition system is γ2.

[0073] In this embodiment, γ1=1.02, γ2=0.97. When the performance index of the automatic surgical procedure recognition system is less than the first threshold value of the performance index of the automatic surgical procedure recognition system and greater than the second threshold value of the performance index of the automatic surgical procedure recognition system, that is, γ2<γ<γ1, it indicates that the performance index of the automatic surgical procedure recognition system is in the normal range, that is, the performance of the automatic surgical procedure recognition system is in the passing range; when in other situations, that is, γ≥γ1 or γ≤γ2, it indicates that the performance index of the automatic surgical procedure recognition system is in the abnormal range, that is, the performance of the automatic surgical procedure recognition system is in the unsatisfactory range.

[0074] Optionally, the feedback module provides feedback on the performance of the automatic surgical procedure recognition system based on video analysis according to the performance indicators of the automatic surgical procedure recognition system. When the performance indicators of the automatic surgical procedure recognition system are in the unsatisfactory range, the recognition results of the automatic surgical procedure recognition system fed back deviate greatly from the simulated surgical operation process; when the performance indicators of the automatic surgical procedure recognition system are in the passing range, the recognition results of the automatic surgical procedure recognition system fed back are more consistent with the simulated surgical operation process.

[0075] In this embodiment, when the performance index of the automatic recognition system of the surgical process is in the qualified range, it indicates that the result recognized by the automatic recognition system of the surgical process is consistent with the actual situation, that is, the performance of the automatic recognition system of the surgical process is good. When the performance index of the automatic recognition system of the surgical process is in the unqualified range, it indicates that the result recognized by the automatic recognition system of the surgical process is greatly different from the actual situation, that is, the performance of the automatic recognition system of the surgical process is poor. When the performance index of the automatic recognition system of the surgical process is in the unqualified range, the automatic recognition system of the surgical process can replace other automatic recognition methods of the surgical process, and mark and record the automatic recognition method of the surgical process whose performance index of the current automatic recognition system of the surgical process is in the unqualified range, and optimize the automatic recognition system and method of the surgical process. At this time, the automatic recognition system of the surgical process can be optimized to specifically include obtaining multiple data sources such as video and sensors, obtaining more comprehensive surgical process information, obtaining clearer video data, more accurate and detailed data annotation of surgical video data, more accurate annotation of each step in the surgical process including transition operations, and collecting more surgical data of different difficulties and different types, increasing the generalization ability of the automatic recognition system of the surgical process, and obtaining a more accurate automatic recognition system of the surgical process, and obtaining a more accurate recognition result.

[0076] In summary, the embodiments of the present invention obtain surgical step accuracy data and surgical process recognition accuracy data to obtain surgical step accuracy indicators and surgical process recognition accuracy indicators, and then obtain surgical process automatic recognition system performance indicators based on the surgical step accuracy indicators and surgical process recognition accuracy indicators, thereby providing feedback on the degree of proximity between the recognition results of the surgical process automatic recognition system and the simulated surgical operation process based on the performance indicators of the surgical process automatic recognition system, thereby achieving accurate analysis of the results of automatic surgical process recognition, and effectively solving the problem of inaccurate analysis of the results of automatic surgical process recognition in the prior art.

[0077] There are a few points to note:

[0078] (1) The drawings of the embodiments of the present invention only relate to the structures related to the embodiments of the present invention, and other structures may refer to the general design.

[0079] (2) For the sake of clarity, in the drawings used to describe the embodiments of the present invention, the thickness of the layers or regions is exaggerated or reduced, that is, these drawings are not drawn according to the actual scale. It is understood that when an element such as a layer, film, region or substrate is referred to as being "on" or "under" another element, the element may be "directly" "on" or "under" the other element or there may be intermediate elements.

[0080] (3) In the absence of conflict, the embodiments of the present invention and the features therein may be combined with each other to obtain new embodiments.

[0081] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. The protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A surgical procedure automatic recognition system based on video analysis, characterized in that: It includes a data acquisition module, a surgical step accuracy index acquisition module, a surgical process recognition accuracy index acquisition module, a data analysis module and a feedback module; Wherein, the data acquisition module is used to acquire the surgical step accuracy data and the surgical process recognition accuracy data; The surgical step accuracy index acquisition module is used to acquire the surgical step accuracy index according to the surgical step accuracy data; The surgical procedure recognition accuracy index acquisition module is used to acquire the surgical procedure recognition accuracy index according to the surgical procedure recognition accuracy data; The data analysis module is used to obtain the performance index of the automatic surgical process recognition system according to the surgical step accuracy index and the surgical process recognition accuracy index; The feedback module is used to feedback the performance of the automatic surgical procedure recognition system according to the performance indicators of the automatic surgical procedure recognition system; The specific method for obtaining the surgical process recognition accuracy index is as follows: Obtain the types of instruments used in the surgical procedure and number the types of instruments used in the surgical procedure; Obtaining from the database the preset standard data of surgical incision length, standard data of surgical time, standard data of surgical blood loss, and the weight ratio of the number of times each instrument is used in the simulated surgical process in the surgical process recognition accuracy index; Construct a calculation formula for surgical procedure recognition accuracy index; The specific calculation formula for the surgical process recognition accuracy index is: In the formula, θ represents the surgical process recognition accuracy index, a represents the actual data of surgical incision length, a1 represents the standard data of surgical incision length, b represents the actual data of surgical time, b1 represents the standard data of surgical time, c represents the actual data of surgical bleeding, c1 represents the standard data of surgical bleeding, and e represents the actual data of surgical bleeding. i It represents the number of times the i-th device identified by the system is used, i represents the number of the device type, i = 1, 2, ..., n, n represents the total number of device types, e i ′ It is represented by the number of times the i-th instrument is used in the simulated surgical process, α is represented by the accuracy index of the surgical steps, d1 is represented by the weight ratio of the actual data of the surgical incision length in the surgical process recognition accuracy index, d2 is represented by the weight ratio of the actual data of the surgical time in the surgical process recognition accuracy index, d3 is represented by the weight ratio of the actual data of the surgical bleeding volume in the surgical process recognition accuracy index, d4 is represented by the weight ratio of the number of times each instrument is used identified by the system in the surgical process recognition accuracy index, and d5 is represented by the weight ratio of the surgical step accuracy index in the surgical process recognition accuracy index.

2. The automatic surgical process recognition system based on video analysis according to claim 1 is characterized in that: The surgical step accuracy data includes surgical step recognition accuracy, surgical step recognition recall rate and surgical step recognition F1 score; The surgical procedure recognition accuracy data includes surgical procedure recognition accuracy standard data and surgical procedure recognition accuracy actual data.

3. The automatic surgical procedure recognition system based on video analysis according to claim 2 is characterized in that: The standard data for surgical process recognition accuracy include standard data for surgical incision length, standard data for surgical time, standard data for surgical blood loss, and the number of times each instrument is used in the simulated surgical process; The actual data on surgical process recognition accuracy includes actual data on surgical incision length, actual data on surgical time, actual data on surgical bleeding volume, and the number of times each instrument is used as identified by the automatic surgical process recognition system based on video analysis.

4. The automatic surgical procedure recognition system based on video analysis according to claim 2, characterized in that: The surgical step recognition accuracy is the ratio of the number of surgical procedure steps correctly recognized by the surgical procedure automatic recognition system to the total number of surgical procedure steps recognized by the surgical procedure automatic recognition system; The surgical step recognition recall rate is the ratio of the number of process steps correctly recognized by the surgical process automatic recognition system to the total number of simulated surgical process steps; The surgical step recognition F1 score is the harmonic mean of the surgical step recognition accuracy and the surgical step recognition recall.

5. The automatic surgical procedure recognition system based on video analysis according to claim 1, characterized in that: The specific analysis process of the surgical step accuracy index is as follows: Obtaining a surgical step accuracy index, obtaining a first threshold value of the surgical step accuracy index and a second threshold value of the surgical step accuracy index, comparing the surgical step accuracy index with the first threshold value of the surgical step accuracy index and the second threshold value of the surgical step accuracy index, when the surgical step accuracy index is less than the first threshold value of the surgical step accuracy index and greater than the second threshold value of the surgical step accuracy index, it indicates that the surgical step accuracy index is within a normal range, that is, the surgical step accuracy index is within a passing range; When in other situations, it means that the surgical step accuracy index is in an abnormal range, that is, the surgical step accuracy index is in a failing range.

6. The automatic surgical procedure recognition system based on video analysis according to claim 1, characterized in that: The specific analysis process of the surgical procedure recognition accuracy index is as follows: Obtaining a surgical procedure recognition accuracy index, obtaining a first threshold value of the surgical procedure recognition accuracy index and a second threshold value of the surgical procedure recognition accuracy index, comparing the surgical procedure recognition accuracy index with the first threshold value of the surgical procedure recognition accuracy index and the second threshold value of the surgical procedure recognition accuracy index, when the surgical procedure recognition accuracy index is less than the first threshold value of the surgical procedure recognition accuracy index and greater than the second threshold value of the surgical procedure recognition accuracy index, it indicates that the surgical procedure recognition accuracy index is within a normal range, that is, the surgical procedure recognition accuracy index is within a passing range; When in other situations, it means that the surgical process recognition accuracy index is in an abnormal range, that is, the surgical process recognition accuracy index is in a failing range.

7. The automatic surgical procedure recognition system based on video analysis according to claim 1, characterized in that: The specific method for obtaining the performance indicators of the surgical procedure automatic identification system is as follows: Wherein, γ represents the performance index of the automatic recognition system for surgical procedures, α represents the accuracy index of surgical steps, θ represents the accuracy index of surgical procedure recognition, w1 represents the weight ratio of the accuracy index of surgical steps in the performance index of the automatic recognition system for surgical procedures, w2 represents the weight ratio of the accuracy index of surgical procedure recognition in the performance index of the automatic recognition system for surgical procedures, and e is a natural constant.

8. The automatic surgical procedure recognition system based on video analysis according to claim 7, characterized in that: The specific analysis process of the performance indicators of the surgical procedure automatic identification system is as follows: Obtaining a performance indicator of a surgical procedure automatic recognition system, obtaining a first threshold value and a second threshold value of the surgical procedure automatic recognition system performance indicator, comparing the surgical procedure automatic recognition system performance indicator with the first threshold value and the second threshold value of the surgical procedure automatic recognition system performance indicator, when the surgical procedure automatic recognition system performance indicator is less than the first threshold value and greater than the second threshold value, it indicates that the surgical procedure automatic recognition system performance indicator is within a normal range, that is, the surgical procedure automatic recognition system performance indicator is within a qualified range; When in other situations, it means that the performance indicators of the automatic recognition system for surgical procedures are in an abnormal range, that is, the performance indicators of the automatic recognition system for surgical procedures are in a failing range.

9. The automatic surgical procedure recognition system based on video analysis according to claim 1, characterized in that: The feedback module provides feedback on the performance of the automatic surgical process recognition system based on video analysis according to the performance index of the automatic surgical process recognition system. When the performance index of the automatic surgical process recognition system is in a failing range, the recognition result of the automatic surgical process recognition system is greatly deviated from the simulated surgical operation process. When the performance indicators of the automatic surgical procedure recognition system are within the passing range, the recognition results fed back by the automatic surgical procedure recognition system are consistent with the simulated surgical operation process.

Images