An endotracheal tube positioning method and system for assisted guidance

Abstract

The application discloses an endotracheal tube positioning method and system for assisted guiding, relates to the field of medical instrument use assistance, and is applied to an endotracheal tube assisted guiding system. The endotracheal tube assisted guiding system comprises a visual laryngoscope, an endotracheal tube and a gas sensor. The visual laryngoscope is electrically connected with the gas sensor, and the gas sensor is installed at the most front end of the endotracheal tube in the insertion direction. The method comprises the following steps: collecting a first time sequence signal for representing the carbon dioxide concentration inside a target airway by the gas sensor, collecting an airway image of the target airway by the visual laryngoscope; performing anatomical structure recognition on the airway image to obtain an anatomical structure image; performing time sequence feature processing on the first time sequence signal, generating a visual mask and superimposing the visual mask on the anatomical structure image to generate a fusion display image; and outputting the positioning information of the endotracheal tube based on the fusion display image and in combination with the visual laryngoscope. Through implementation of the application, the catheter position can be accurately positioned in the intubation process.

Description

Technical Field

[0001] This application relates to the field of assistive medical device use, and more particularly to a method and system for assisted guidance of airway catheter positioning. Background Technology

[0002] In emergency clinical management, intubation of difficult airways is characterized by high risk and a high complication rate, and initial intubation failure can easily lead to patient hypoxia, brain injury, or even death. Therefore, it is necessary to optimize the difficult airway intubation procedure, provide guidance and assistance during the process, and improve the efficiency and success rate of airway intubation.

[0003] Currently, clinical airway intubation typically uses a video laryngoscope, relying on visual aids for judgment. However, this method heavily depends on the operator's subjective judgment. In emergency situations with abundant secretions or anatomical variations, subjective judgment is insufficient to determine the catheter's location, easily leading to errors such as misinsertion of the esophagus or repeated attempts, failing to meet the actual needs of assisted guidance. While AI-assisted intubation technology has developed alongside artificial intelligence, current AI-based intubation assistance, based on visual image recognition, is susceptible to interference from complex clinical environments such as obstructed view, insufficient lighting, and mucosal bleeding. This limits the accuracy of the assessment and still fails to meet the actual needs of assisted guidance. Therefore, accurately locating the catheter to accurately assist and guide difficult airway intubation remains a critical technical problem that needs to be solved. Summary of the Invention

[0004] This application provides a method and system for assisted airway catheter positioning to solve the technical problem of inaccurate catheter positioning during existing intubation processes.

[0005] According to a first aspect of the embodiments of this application, a method for positioning an endotracheal tube for assisted guidance is provided, applied to an endotracheal tube assisted guidance system. The endotracheal tube assisted guidance system includes a video laryngoscope, an endotracheal tube, and a gas sensor. The video laryngoscope is electrically connected to the gas sensor, and the gas sensor is installed at the foremost end of the endotracheal tube in the direction of insertion. The method includes: A first time-series signal characterizing carbon dioxide concentration inside the target airway and an airway image of the target airway are acquired; wherein, the first time-series signal is acquired by the gas sensor; and the airway image is acquired by the video laryngoscope. Anatomical structure recognition is performed on the airway image to obtain the anatomical structure image of the target airway; The first time-series signal is subjected to time-series feature processing to generate a visual mask for characterizing changes in carbon dioxide concentration, and the visual mask is superimposed on the anatomical structure image to generate a fused display image; Based on the fused display image and combined with the visual laryngoscope, the positioning information of the endotracheal tube, corresponding to the positional relationship between the endotracheal tube and the target airway, is output.

[0006] This application is based on an endotracheal tube-assisted guidance system including a video laryngoscope, an endotracheal tube, and a gas sensor. First, a first temporal signal representing carbon dioxide concentration inside the target airway is acquired using the gas sensor. Then, an airway image of the target airway is acquired using the video laryngoscope. Next, anatomical structure recognition is performed on the airway image to obtain an anatomical structure image of the target airway. Then, temporal feature processing is performed on the first temporal signal to generate a visual mask representing changes in carbon dioxide concentration, which is superimposed on the anatomical structure image to obtain a fused display image. Finally, based on the fused display image and combined with the video laryngoscope, the positioning information of the endotracheal tube corresponding to the positional relationship between the endotracheal tube and the target airway is output. By using a gas sensor installed at the very tip of the endotracheal tube in the direction of entry, more accurate acquisition of... The carbon dioxide concentration within the target airway, and the resulting visual mask, can more accurately represent changes in carbon dioxide concentration. These changes, in turn, more accurately characterize the positional relationship between the endotracheal tube and the target airway, thus accurately locating the endotracheal tube. By acquiring airway images of the target airway through a video laryngoscope and performing anatomical structure recognition to obtain anatomical structure images, which are then superimposed on the visual mask, compared to existing methods that rely solely on video laryngoscope visualization or visual image recognition, this method, by performing anatomical structure recognition on airway images and combining it with a carbon dioxide concentration-related visual mask for tube location, reduces reliance on the operator's subjective judgment and avoids interference from difficult visual environments such as visual field obstruction, thereby improving the accuracy of endotracheal tube location.

[0007] In some embodiments of this application, the step of performing anatomical structure recognition on the airway image to obtain an anatomical structure image of the target airway specifically includes: Based on a pre-trained convolutional neural network, anatomical structure recognition is performed on the airway image to obtain the anatomical structure image of the target airway.

[0008] This application uses a pre-trained convolutional neural network to identify the anatomical structure of airway images, which can accurately extract the visual features of the target airway corresponding to the airway image. Then, when locating the endotracheal tube based on the obtained anatomical structure image and a visual mask, it can reduce the dependence on the operator's subjective judgment and avoid interference from difficult visual environments such as visual field obstruction, thereby improving the accuracy of endotracheal tube location.

[0009] In some embodiments of this application, the step of performing time-series feature processing on the first time-series signal to generate a visual mask for characterizing changes in carbon dioxide concentration specifically includes: Based on a pre-trained temporal neural network, the first temporal signal is processed to obtain the carbon dioxide concentration change characteristics of the endotracheal tube located inside the target airway, and a visual mask for characterizing the carbon dioxide concentration change is generated based on the carbon dioxide concentration change characteristics.

[0010] This application uses a pre-trained temporal neural network to perform temporal feature processing on the first temporal signal, which can fully explore the carbon dioxide concentration change features related to the change in the position of the endotracheal tube inside the target airway. The resulting visual mask can accurately represent the carbon dioxide concentration change. Therefore, when locating the tube in combination with anatomical images, it can reduce the dependence on the operator's subjective judgment and avoid interference from difficult visual environments such as visual field obstruction, thereby improving the accuracy of endotracheal tube location.

[0011] In some embodiments of this application, the step of outputting the positioning information of the endotracheal tube, corresponding to the positional relationship between the endotracheal tube and the target airway, based on the fused display image and in conjunction with the video laryngoscope, specifically includes: Based on a pre-trained localization neural network, the fused display image is identified to obtain the position guidance information of the endotracheal tube; Based on the fused display image and the visual mask, the positional relationship information between the endotracheal tube and the target airway is determined; Using the video laryngoscope, the positional relationship between the endotracheal tube and the target airway is determined by combining the positioning guidance information and the positional relationship information, and the positioning information of the endotracheal tube corresponding to the positional relationship is output.

[0012] This application first uses a pre-trained localization neural network to identify and fuse the displayed image to obtain the position guidance information of the endotracheal tube. It can fully identify the spatial relationship between anatomical structures and visual masks in the fused displayed image. Then, through the fused displayed image and visual mask, it determines the positional relationship information between the endotracheal tube and the target airway. Based on the carbon dioxide concentration change represented by the visual mask, it can objectively verify and identify the positional relationship between the endotracheal tube and the target airway. Then, through a video laryngoscope, the positional relationship between the endotracheal tube and the target airway is determined by combining the localization guidance information and the positional relationship information. Through the mutual supplementation and cross-validation of the localization guidance information and the positional relationship information, an accurate positional relationship can be obtained, reducing the risk of mispositioning, reducing the dependence on the operator's subjective judgment, and avoiding interference in difficult visual environments such as visual field obstruction, thereby improving the accuracy of endotracheal tube location.

[0013] In some embodiments of this application, the endotracheal tube-assisted guidance system further includes an audible and visual cue module; before acquiring the first time-series signal inside the target airway used to characterize carbon dioxide concentration and the airway image of the target airway, the system further includes: The gas sensor is used to acquire a second time-series signal that characterizes the carbon dioxide concentration in the current channel where the endotracheal tube is located; Based on the second timing signal and combined with the preset carbon dioxide concentration threshold, the relationship between the current channel and the target airway and the position status of the endotracheal tube are determined, and the corresponding audio-visual prompt information is output through the audio-visual prompt module.

[0014] Before acquiring the first time-series signal and airway image, this application first uses a gas sensor to collect a second time-series signal representing the carbon dioxide concentration in the current channel where the endotracheal tube is located. Then, combined with a preset carbon dioxide concentration threshold, it determines the relationship between the current channel and the target airway, as well as the position and status of the endotracheal tube. This allows for rapid and objective identification of whether the endotracheal tube has mistakenly entered a non-target channel in the early stages of endotracheal tube insertion, thereby accurately locating the endotracheal tube. Subsequently, the application outputs corresponding audio-visual prompts through an audio-visual prompt module, which can further prevent invalid insertion of the endotracheal tube, improve the accuracy of endotracheal tube location, and avoid the risks caused by invalid endotracheal tube insertion.

[0015] In some embodiments of this application, the step of determining the relationship between the current channel and the target airway and the positional state of the endotracheal tube based on the second timing signal and a preset carbon dioxide concentration threshold, and outputting corresponding audio-visual prompt information through the audio-visual prompt module, specifically includes: If the carbon dioxide concentration corresponding to the second timing signal is lower than the preset first carbon dioxide concentration threshold, it is determined that the current channel is different from the target airway and the position of the tracheal tube is abnormal. The corresponding warning audio-visual prompt is output through the audio-visual prompt module. If the carbon dioxide concentration corresponding to the second timing signal is not lower than the first carbon dioxide concentration threshold and is lower than the preset second carbon dioxide concentration threshold, then it is determined that the current channel is the same as the target airway and the position of the tracheal tube is abnormal. The corresponding abnormal audio-visual prompt is output through the audio-visual prompt module. If the carbon dioxide concentration corresponding to the second timing signal is not lower than the second carbon dioxide concentration threshold, then it is determined that the current channel is the same as the target airway and the position of the tracheal tube is normal. The corresponding normal audio-visual prompt is output through the audio-visual prompt module, and the second timing signal is used as the first timing signal.

[0016] This application, through an objective comparison of the carbon dioxide concentration corresponding to the second time-series signal with the first and second carbon dioxide concentration thresholds, further determines the relationship between the current channel and the target airway, as well as the position and status of the endotracheal tube. It can objectively classify different states of the endotracheal tube through a carbon dioxide concentration grading mechanism, quickly and objectively identifying whether the endotracheal tube has been mistakenly inserted into a non-target channel in the early stages of insertion. This allows for accurate positioning of the endotracheal tube, followed by the output of corresponding audio-visual prompts via an audio-visual prompt module. This further avoids invalid insertion of the endotracheal tube and provides differentiated guidance, improving the accuracy of endotracheal tube positioning while avoiding the risks caused by invalid insertion.

[0017] According to a second aspect of the embodiments of this application, an endotracheal tube positioning system for assisted guidance is provided, applied to an endotracheal tube assisted guidance system. The endotracheal tube assisted guidance system includes a video laryngoscope, an endotracheal tube, and a gas sensor. The video laryngoscope is electrically connected to the gas sensor, and the gas sensor is installed at the foremost end of the endotracheal tube in the direction of entry. The endotracheal tube positioning system includes a data acquisition module, a first processing module, a second processing module, and a tube positioning module. The data acquisition module is used to acquire a first time-series signal representing carbon dioxide concentration inside the target airway and an airway image of the target airway; wherein, the first time-series signal is acquired by the gas sensor; and the airway image is acquired by the video laryngoscope. The first processing module is used to perform anatomical structure recognition on the airway image to obtain an anatomical structure image of the target airway; The second processing module is used to perform time-series feature processing on the first time-series signal, generate a visual mask for characterizing changes in carbon dioxide concentration, and superimpose the visual mask onto the anatomical structure image to generate a fused display image; The endotracheal tube positioning module is used to output positioning information of the endotracheal tube corresponding to the positional relationship between the endotracheal tube and the target airway, based on the fused display image and in conjunction with the visual laryngoscope.

[0018] In some embodiments of this application, the first processing module includes a first processing unit; the first processing unit is used to perform anatomical structure recognition on the airway image based on a pre-trained convolutional neural network to obtain an anatomical structure image of the target airway.

[0019] In some embodiments of this application, the second processing module includes a second processing unit; the second processing unit is used to perform time-series feature processing on the first time-series signal based on a pre-trained time-series neural network to obtain the carbon dioxide concentration change characteristics of the tracheal tube located inside the target airway, and generate a visual mask for characterizing the carbon dioxide concentration change based on the carbon dioxide concentration change characteristics.

[0020] In some embodiments of this application, the catheter positioning module includes a first information determination unit, a second information determination unit, and a positioning information output unit; The first information determining unit is used to identify the fused display image based on a pre-trained positioning neural network to obtain the position guidance information of the endotracheal tube; The second information determining unit is used to determine the positional relationship information between the endotracheal tube and the target airway based on the fused display image and in conjunction with the visual mask; The positioning information output unit is used to output the positioning information of the endotracheal tube corresponding to the positional relationship between the endotracheal tube and the target airway, determined by the video laryngoscope in combination with the positioning guidance information and the positional relationship information.

[0021] In some embodiments of this application, the endotracheal tube-assisted guidance system further includes an audible and visual cue module; before acquiring the time-series signal of carbon dioxide concentration inside the target airway and the airway image of the target airway, the system further includes: The gas sensor is used to acquire a second time-series signal that characterizes the carbon dioxide concentration in the current channel where the endotracheal tube is located; Based on the second timing signal and combined with the preset carbon dioxide concentration threshold, the relationship between the current channel and the target airway and the position status of the endotracheal tube are determined, and the corresponding audio-visual prompt information is output through the audio-visual prompt module.

[0022] In some embodiments of this application, the step of determining the relationship between the current channel and the target airway and the positional state of the endotracheal tube based on the second timing signal and a preset carbon dioxide concentration threshold, and outputting corresponding audio-visual prompt information through the audio-visual prompt module, specifically includes: If the carbon dioxide concentration corresponding to the second timing signal is lower than the preset first carbon dioxide concentration threshold, it is determined that the current channel is different from the target airway and the position of the tracheal tube is abnormal. The corresponding warning audio-visual prompt is output through the audio-visual prompt module. If the carbon dioxide concentration corresponding to the second timing signal is not lower than the first carbon dioxide concentration threshold and is lower than the preset second carbon dioxide concentration threshold, then it is determined that the current channel is the same as the target airway and the position of the tracheal tube is abnormal. The corresponding abnormal audio-visual prompt is output through the audio-visual prompt module. If the carbon dioxide concentration corresponding to the second timing signal is not lower than the second carbon dioxide concentration threshold, then it is determined that the current channel is the same as the target airway and the position of the tracheal tube is normal. The corresponding normal audio-visual prompt is output through the audio-visual prompt module, and the second timing signal is used as the first timing signal.

[0023] This application is based on an endotracheal tube-assisted guidance system including a video laryngoscope, an endotracheal tube, and a gas sensor. First, a first temporal signal representing carbon dioxide concentration inside the target airway is acquired using the gas sensor. Then, an airway image of the target airway is acquired using the video laryngoscope. Next, anatomical structure recognition is performed on the airway image to obtain an anatomical structure image of the target airway. Then, temporal feature processing is performed on the first temporal signal to generate a visual mask representing changes in carbon dioxide concentration, which is superimposed on the anatomical structure image to obtain a fused display image. Finally, based on the fused display image and combined with the video laryngoscope, the positioning information of the endotracheal tube corresponding to the positional relationship between the endotracheal tube and the target airway is output. By using a gas sensor installed at the very tip of the endotracheal tube in the direction of entry, more accurate acquisition of... The carbon dioxide concentration within the target airway, and the resulting visual mask, can more accurately represent changes in carbon dioxide concentration. These changes, in turn, more accurately characterize the positional relationship between the endotracheal tube and the target airway, thus accurately locating the endotracheal tube. By acquiring airway images of the target airway through a video laryngoscope and performing anatomical structure recognition to obtain anatomical structure images, which are then superimposed on the visual mask, compared to existing methods that rely solely on video laryngoscope visualization or visual image recognition, this method, by performing anatomical structure recognition on airway images and combining it with a carbon dioxide concentration-related visual mask for tube location, reduces reliance on the operator's subjective judgment and avoids interference from difficult visual environments such as visual field obstruction, thereby improving the accuracy of endotracheal tube location. Attached Figure Description

[0024] Figure 1 This is a schematic flowchart illustrating a method for positioning an airway catheter for assisted guidance, as shown in certain embodiments of this application. Figure 2 This is a connection diagram of an endotracheal tube-assisted guidance system shown in certain embodiments of this application. Figure 3 This is a modular structure diagram of an airway catheter positioning system for assisted guidance, as shown in certain embodiments of this application. Detailed Implementation

[0025] The embodiments of this application are described in detail below. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below in conjunction with the accompanying drawings are exemplary and are only used to explain some embodiments of this application, and should not be construed as limiting the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments shown in this application without inventive effort are within the protection scope of this application.

[0026] In the description of this application, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, unless otherwise explicitly specified, "a plurality of" or "several" means two or more.

[0027] Currently, clinical airway intubation typically uses a video laryngoscope, relying on visual aids for judgment. However, this method heavily depends on the operator's subjective judgment. In emergency situations with abundant secretions or anatomical variations, subjective judgment is insufficient to determine the catheter's location, easily leading to errors such as misinsertion of the esophagus or repeated attempts, failing to meet the actual needs of assisted guidance. While AI-assisted intubation technology has developed alongside artificial intelligence, current AI-based intubation assistance, based on visual image recognition, is susceptible to interference from complex clinical environments such as obstructed view, insufficient lighting, and mucosal bleeding. This limits the accuracy of the assessment and still fails to meet the actual needs of assisted guidance. Therefore, accurately locating the catheter to accurately assist and guide difficult airway intubation remains a critical technical problem that needs to be solved.

[0028] Based on the above technical background, please refer to Figure 1 This application provides a method for assisted airway catheter positioning, applied to an endotracheal catheter-assisted guidance system. The endotracheal catheter-assisted guidance system can automatically execute the endotracheal catheter positioning method provided in this application. Please refer to... Figure 2The endotracheal tube-assisted guidance system includes a video laryngoscope 21, an endotracheal tube 22, a gas sensor 23, and an audio-visual prompting module 24. The video laryngoscope 21 includes an image acquisition end 211 and an image display end. The video laryngoscope 21 is electrically connected to the gas sensor 23 via the audio-visual prompting module 24. The video laryngoscope 21 and the audio-visual prompting module 24, and the audio-visual prompting module 24 and the gas sensor 23 are preferably connected using flexible signal lines. The direction of the arrow along the endotracheal tube 22 is the insertion direction of the endotracheal tube 22. The gas sensor 23 is installed at the foremost end of the insertion direction of the endotracheal tube 22.

[0029] Please continue reading Figure 1 The method for assisted airway catheter positioning provided in this application includes steps S101 to S104, each step of which is as follows: Step S101: Acquire a first time-series signal inside the target airway to characterize the carbon dioxide concentration and an airway image of the target airway; wherein, the first time-series signal is acquired by a gas sensor 23; and the airway image is acquired by the acquisition end 211 of a video laryngoscope 21.

[0030] In some embodiments of this application, before acquiring the first time-series signal inside the target airway used to characterize the carbon dioxide concentration and the airway image of the target airway, the method further includes: The gas sensor 23 acquires a second time-series signal that characterizes the carbon dioxide concentration in the current channel where the endotracheal tube 22 is located; Based on the second timing signal and combined with the preset carbon dioxide concentration threshold, the relationship between the current channel and the target airway and the position status of the endotracheal tube are determined, and the corresponding audio-visual prompt information is output through the audio-visual prompt module 24.

[0031] Before acquiring the first time-series signal and airway image, this application first uses a gas sensor to collect a second time-series signal representing the carbon dioxide concentration in the current channel where the endotracheal tube is located. Then, combined with a preset carbon dioxide concentration threshold, it determines the relationship between the current channel and the target airway, as well as the position and status of the endotracheal tube. This allows for rapid and objective identification of whether the endotracheal tube has mistakenly entered a non-target channel in the early stages of endotracheal tube insertion, thereby accurately locating the endotracheal tube. Subsequently, the application outputs corresponding audio-visual prompts through an audio-visual prompt module, which can further prevent invalid insertion of the endotracheal tube, improve the accuracy of endotracheal tube location, and avoid the risks caused by invalid endotracheal tube insertion.

[0032] In some embodiments of this application, the step of determining the relationship between the current channel and the target airway and the position status of the endotracheal tube based on the second timing signal and a preset carbon dioxide concentration threshold, and outputting corresponding audio-visual prompt information through the audio-visual prompt module 24, specifically includes: If the carbon dioxide concentration corresponding to the second timing signal is lower than the preset first carbon dioxide concentration threshold, it is determined that the current channel is different from the target airway and the position of the tracheal tube is abnormal. The corresponding warning audio-visual prompt is output through the audio-visual prompt module 24. If the carbon dioxide concentration corresponding to the second timing signal is not lower than the first carbon dioxide concentration threshold and is lower than the preset second carbon dioxide concentration threshold, then it is determined that the current channel is the same as the target airway and the position of the tracheal tube is abnormal. The corresponding abnormal sound and light prompt is output through the sound and light prompt module 24. If the carbon dioxide concentration corresponding to the second timing signal is not lower than the second carbon dioxide concentration threshold, then it is determined that the current channel is the same as the target airway and the position of the tracheal tube is normal. The corresponding normal audio-visual prompt is output through the audio-visual prompt module 24, and the second timing signal is used as the first timing signal.

[0033] Generally, during intubation, depending on whether the tube is in the correct position, it can be divided into correct endotracheal intubation and incorrect esophageal intubation. It is known to those skilled in the art that the carbon dioxide concentration in the trachea and esophagus is not the same, and the carbon dioxide concentration in the trachea is usually higher. Therefore, endotracheal intubation and esophageal intubation can be distinguished by setting specific thresholds.

[0034] Specifically, the preferred value for the first carbon dioxide concentration threshold is 10 mmHg, and the preferred value for the second carbon dioxide concentration threshold is 30 mmHg. When the carbon dioxide concentration corresponding to the second timing signal is lower than the first carbon dioxide concentration threshold of 10 mmHg, it indicates that the catheter is currently in the esophagus, which is an incorrect insertion. A timely warning signal should be issued via the audible and visual prompt module, such as flashing red lights and continuous beeping, to indicate that the esophagus has been incorrectly inserted and that re-insertion is necessary to avoid further catheterization that could delay rescue efforts and cause secondary injuries. When the carbon dioxide concentration corresponding to the second timing signal is not lower than the second carbon dioxide concentration threshold of 30 mmHg... If the current catheter is in the correct position in the trachea, indicating correct intubation, a normal audible and visual prompt should be issued via the audible and visual prompt module, such as a solid green light and a short, continuous single beep, to indicate that the intubation position is correct and the catheter can continue to be inserted or fixed. If the carbon dioxide concentration corresponding to the second time sequence signal is between the first carbon dioxide concentration threshold of 10 mmHg and the second carbon dioxide concentration threshold of 30 mmHg, it indicates that the current catheter is in the trachea but not in the correct position, indicating abnormal intubation. An abnormal audible and visual prompt should be issued via the audible and visual prompt module, such as a flashing yellow light and an intermittent beep, to indicate that the intubation position is abnormal and adjustment is required.

[0035] This application, through an objective comparison of the carbon dioxide concentration corresponding to the second time-series signal with the first and second carbon dioxide concentration thresholds, further determines the relationship between the current channel and the target airway, as well as the position and status of the endotracheal tube. It can objectively classify different states of the endotracheal tube through a carbon dioxide concentration grading mechanism, quickly and objectively identifying whether the endotracheal tube has been mistakenly inserted into a non-target channel in the early stages of insertion. This allows for accurate positioning of the endotracheal tube, followed by the output of corresponding audio-visual prompts via an audio-visual prompt module. This further avoids invalid insertion of the endotracheal tube and provides differentiated guidance, improving the accuracy of endotracheal tube positioning while avoiding the risks caused by invalid insertion.

[0036] Step S102: Perform anatomical structure recognition on the airway image to obtain the anatomical structure image of the target airway.

[0037] In some embodiments of this application, the step of performing anatomical structure recognition on the airway image to obtain an anatomical structure image of the target airway specifically includes: Based on a pre-trained convolutional neural network, anatomical structure recognition is performed on the airway image to obtain the anatomical structure image of the target airway.

[0038] Specifically, the convolutional neural network used for anatomical structure recognition includes, but is not limited to, CNN and its improved variants, with CNN being the preferred implementation.

[0039] This application uses a pre-trained convolutional neural network to identify the anatomical structure of airway images, which can accurately extract the visual features of the target airway corresponding to the airway image. Then, when locating the endotracheal tube based on the obtained anatomical structure image and a visual mask, it can reduce the dependence on the operator's subjective judgment and avoid interference from difficult visual environments such as visual field obstruction, thereby improving the accuracy of endotracheal tube location.

[0040] Step S103: Perform time-series feature processing on the first time-series signal to generate a visual mask for characterizing changes in carbon dioxide concentration, and superimpose the visual mask onto the anatomical structure image to generate a fused display image.

[0041] In some embodiments of this application, the step of performing time-series feature processing on the first time-series signal to generate a visual mask for characterizing changes in carbon dioxide concentration specifically includes: Based on a pre-trained temporal neural network, the first temporal signal is processed to obtain the carbon dioxide concentration change characteristics of the endotracheal tube located inside the target airway, and a visual mask for characterizing the carbon dioxide concentration change is generated based on the carbon dioxide concentration change characteristics.

[0042] Specifically, the temporal neural network used for temporal feature processing includes, but is not limited to, LSTM and its improved variants, with LSTM being the preferred implementation.

[0043] Specifically, the generated visual mask is preferably in the form of a semi-transparent thermal image to facilitate overlay with the anatomical structure image. When generating the visual mask, the corresponding carbon dioxide concentration change features can be mapped to a blue grayscale gradient. Then, by changing the fill color of the visual mask accordingly, the corresponding carbon dioxide concentration change can be characterized. The preferred blue grayscale gradient is: colorless when below 10 mmHg; light blue when 10-30 mmHg; and dark blue when above 30 mmHg.

[0044] This application uses a pre-trained temporal neural network to perform temporal feature processing on the first temporal signal, which can fully explore the carbon dioxide concentration change features related to the change in the position of the endotracheal tube inside the target airway. The resulting visual mask can accurately represent the carbon dioxide concentration change. Therefore, when locating the tube in combination with anatomical images, it can reduce the dependence on the operator's subjective judgment and avoid interference from difficult visual environments such as visual field obstruction, thereby improving the accuracy of endotracheal tube location.

[0045] Step S104: Based on the fused display image and combined with the visual laryngoscope, output the positioning information of the endotracheal tube 22 corresponding to the positional relationship between the endotracheal tube 22 and the target airway.

[0046] In some embodiments of this application, the step of outputting positioning information of the endotracheal tube 22 corresponding to the positional relationship between the endotracheal tube 22 and the target airway based on the fused display image and in conjunction with the video laryngoscope specifically includes: Based on a pre-trained localization neural network, the fused display image is identified to obtain the position guidance information of the endotracheal tube; Based on the fused display image and the visual mask, the positional relationship information between the endotracheal tube 22 and the target airway is determined; Using the video laryngoscope, the positional relationship between the endotracheal tube 22 and the target airway is determined by combining the positioning guidance information and the positional relationship information, and the positioning information of the endotracheal tube 22 corresponding to the positional relationship is output.

[0047] Specifically, the position guidance information includes the alignment deviation between the catheter and the target position within the target airway and the angle adjustment value of the catheter; the positioning neural network used to identify the fused display image to obtain the position guidance information includes, but is not limited to, CNN, R-CNN, YOLO and corresponding improved variants, or large language models, as long as it can output position guidance information, it can meet the requirements of the positioning neural network.

[0048] Specifically, when determining the positional relationship between the endotracheal tube and the target airway, the positional relationship between the endotracheal tube and the target position within the target airway can be determined by color changes based on the presentation of the visual mask on the fused display image, indicating whether the endotracheal tube is in place or not.

[0049] Specifically, when outputting the positioning information of the endotracheal tube corresponding to the positional relationship, one can choose to directly splice the positional relationship information and the positioning guidance information, and output the positioning information through the audio-visual prompt module 24 or the video laryngoscope 21 in the form of voice broadcast or image display.

[0050] This application first uses a pre-trained localization neural network to identify and fuse the displayed image to obtain the position guidance information of the endotracheal tube. It can fully identify the spatial relationship between anatomical structures and visual masks in the fused displayed image. Then, through the fused displayed image and visual mask, it determines the positional relationship information between the endotracheal tube and the target airway. Based on the carbon dioxide concentration change represented by the visual mask, it can objectively verify and identify the positional relationship between the endotracheal tube and the target airway. Then, through a video laryngoscope, the positional relationship between the endotracheal tube and the target airway is determined by combining the localization guidance information and the positional relationship information. Through the mutual supplementation and cross-validation of the localization guidance information and the positional relationship information, an accurate positional relationship can be obtained, reducing the risk of mispositioning, reducing the dependence on the operator's subjective judgment, and avoiding interference in difficult visual environments such as visual field obstruction, thereby improving the accuracy of endotracheal tube location.

[0051] For specific explanation Figure 2 The implementation process of the endotracheal tube-assisted guidance system for the endotracheal tube positioning method provided in this application is further explained below using the workflow of the endotracheal tube-assisted guidance system: (1) Equipment self-test: Start the endotracheal tube, calibrate the baseline concentration of the gas sensor in the endotracheal tube, and transmit the test signal to the audio-visual prompt module through the flexible signal line inside the tube to verify the autonomous judgment function of the audio-visual prompt module; start the video laryngoscope to verify the imaging clarity of the acquisition end; test the connection stability and signal stability of each flexible signal line used for electrical connection; (2) Set usage mode: (2-1) Mode 1: This mode is the independent use mode of the endotracheal tube. It only relies on the endotracheal tube, its gas sensor, and the audio-visual prompt module to work. It only uses the time-series signal collected by the gas sensor to characterize the carbon dioxide concentration, combined with the carbon dioxide concentration threshold, to determine whether the intubation position is in the trachea (above 30 mmHg), esophagus (below 10 mmHg), or abnormal position (10-30 mmHg), and issues corresponding audio-visual prompts through the audio-visual prompt module. (2-2) Mode 2: This mode is a linkage mode in which the endotracheal tube, its gas sensor, the audio-visual prompt module, and the video laryngoscope all work and execute the endotracheal tube positioning method for auxiliary guidance provided in this application during use. (3) Catheter positioning during intubation: (3-1) Mode 1: The gas sensor collects the time-series signal used to characterize the carbon dioxide concentration during the catheter insertion process. The corresponding carbon dioxide concentration is as follows: when it is below 10 mmHg, it is determined that the catheter is in the esophagus, and the audible and visual prompt module emits a warning audible and visual prompt with flashing red light and continuous beeping; when it is above 30 mmHg, it is determined that the catheter is in the trachea, and the audible and visual prompt module emits a normal audible and visual prompt with a solid green light and a single short continuous beeping sound; when it is between 10-30 mmHg, it is determined that the catheter position is abnormal, and the audible and visual prompt module emits an abnormal audible and visual prompt with flashing yellow light and intermittent beeping. (3-2) Mode 2: Execute the endotracheal tube positioning method for assisted guidance provided in this application. In the early stage of tube intubation, when the carbon dioxide concentration corresponding to the time-series signal collected by the gas sensor is lower than 10 mmHg, a warning audio-visual prompt is issued by the audio-visual prompt module, with flashing red light and continuous beeping, to prompt re-intubation until the carbon dioxide concentration is not lower than 10 mmHg. In the later stage of tube intubation, the carbon dioxide time-series signal and airway image are continuously collected, processed and superimposed into a fused display image. The positional relationship information and positioning guidance information obtained from the fused display image are further determined and output through the audio-visual prompt module or video laryngoscope to provide corresponding operation prompts to the operator. During this process, the audio-visual prompt module continues to use the carbon dioxide concentration threshold (30 mmHg) to assist in judging the tube's positioning.

[0052] Compared to existing technologies, this application is based on an endotracheal tube-assisted guidance system including a video laryngoscope, an endotracheal tube, and a gas sensor. First, a first temporal signal characterizing carbon dioxide concentration inside the target airway is acquired using the gas sensor. Then, an airway image of the target airway is acquired using the video laryngoscope. Next, anatomical structure recognition is performed on the airway image to obtain an anatomical structure image of the target airway. Then, temporal feature processing is performed on the first temporal signal to generate a visual mask characterizing changes in carbon dioxide concentration, which is superimposed on the anatomical structure image to obtain a fused display image. Finally, based on the fused display image and combined with the video laryngoscope, the positioning information of the endotracheal tube corresponding to its positional relationship with the target airway is output. The gas sensor installed at the very tip of the endotracheal tube in the direction of insertion enables more accurate positioning. Accurately obtaining the carbon dioxide concentration within the target airway allows for a more precise visual mask that characterizes changes in carbon dioxide concentration. These changes, in turn, more accurately represent the positional relationship between the endotracheal tube and the target airway, thus accurately locating the endotracheal tube. Furthermore, acquiring airway images of the target airway using a video laryngoscope and performing anatomical structure recognition yields an anatomical structure image, which is then overlaid on the visual mask. Compared to existing methods that rely solely on video laryngoscope visualization or visual image recognition, this approach, which combines anatomical structure recognition of the airway image with a carbon dioxide concentration-related visual mask for tube location, reduces reliance on operator subjective judgment and avoids interference from challenging visual environments such as field-of-view obstruction, thereby improving the accuracy of endotracheal tube location.

[0053] For a method corresponding to the one described above, please refer to [link to relevant documentation]. Figure 3This application provides an airway catheter positioning system for assisted guidance, applied to an endotracheal catheter assisted guidance system. The endotracheal catheter assisted guidance system includes a video laryngoscope, an endotracheal tube, and a gas sensor. The video laryngoscope is electrically connected to the gas sensor, and the gas sensor is installed at the foremost end of the endotracheal tube in the direction of entry. The endotracheal catheter positioning system includes a data acquisition module 310, a first processing module 320, a second processing module 330, and a catheter positioning module 340. The data acquisition module 310 is used to acquire a first time-series signal inside the target airway to characterize the carbon dioxide concentration and an airway image of the target airway; wherein, the first time-series signal is acquired by the gas sensor; and the airway image is acquired by the video laryngoscope. The first processing module 320 is used to perform anatomical structure recognition on the airway image to obtain an anatomical structure image of the target airway; The second processing module 330 is used to perform time-series feature processing on the first time-series signal, generate a visual mask for characterizing changes in carbon dioxide concentration, and superimpose the visual mask onto the anatomical structure image to generate a fused display image. The endotracheal tube positioning module 340 is used to output the positioning information of the endotracheal tube, corresponding to the positional relationship between the endotracheal tube and the target airway, based on the fused display image and in conjunction with the visual laryngoscope.

[0054] In some embodiments of this application, the first processing module 320 includes a first processing unit; the first processing unit is used to perform anatomical structure recognition on the airway image based on a pre-trained convolutional neural network to obtain an anatomical structure image of the target airway.

[0055] In some embodiments of this application, the second processing module 330 includes a second processing unit; the second processing unit is used to perform time-series feature processing on the first time-series signal based on a pre-trained time-series neural network to obtain the carbon dioxide concentration change characteristics of the tracheal tube located inside the target airway, and generate a visual mask for characterizing the carbon dioxide concentration change based on the carbon dioxide concentration change characteristics.

[0056] In some embodiments of this application, the catheter positioning module 340 includes a first information determination unit, a second information determination unit, and a positioning information output unit; The first information determining unit is used to identify the fused display image based on a pre-trained positioning neural network to obtain the position guidance information of the endotracheal tube; The second information determining unit is used to determine the positional relationship information between the endotracheal tube and the target airway based on the fused display image and in conjunction with the visual mask; The positioning information output unit is used to output the positioning information of the endotracheal tube corresponding to the positional relationship between the endotracheal tube and the target airway, determined by the video laryngoscope in combination with the positioning guidance information and the positional relationship information.

[0057] In some embodiments of this application, the endotracheal tube-assisted guidance system further includes an audible and visual cue module; before acquiring the time-series signal of carbon dioxide concentration inside the target airway and the airway image of the target airway, the system further includes: The gas sensor is used to acquire a second time-series signal that characterizes the carbon dioxide concentration in the current channel where the endotracheal tube is located; Based on the second timing signal and combined with the preset carbon dioxide concentration threshold, the relationship between the current channel and the target airway and the position status of the endotracheal tube are determined, and the corresponding audio-visual prompt information is output through the audio-visual prompt module.

[0058] In some embodiments of this application, the step of determining the relationship between the current channel and the target airway and the positional state of the endotracheal tube based on the second timing signal and a preset carbon dioxide concentration threshold, and outputting corresponding audio-visual prompt information through the audio-visual prompt module, specifically includes: If the carbon dioxide concentration corresponding to the second timing signal is lower than the preset first carbon dioxide concentration threshold, it is determined that the current channel is different from the target airway and the position of the tracheal tube is abnormal. The corresponding warning audio-visual prompt is output through the audio-visual prompt module. If the carbon dioxide concentration corresponding to the second timing signal is not lower than the first carbon dioxide concentration threshold and is lower than the preset second carbon dioxide concentration threshold, then it is determined that the current channel is the same as the target airway and the position of the tracheal tube is abnormal. The corresponding abnormal audio-visual prompt is output through the audio-visual prompt module. If the carbon dioxide concentration corresponding to the second timing signal is not lower than the second carbon dioxide concentration threshold, then it is determined that the current channel is the same as the target airway and the position of the tracheal tube is normal. The corresponding normal audio-visual prompt is output through the audio-visual prompt module, and the second timing signal is used as the first timing signal.

[0059] This application is based on an endotracheal tube-assisted guidance system including a video laryngoscope, an endotracheal tube, and a gas sensor. First, a first temporal signal representing carbon dioxide concentration inside the target airway is acquired using the gas sensor. Then, an airway image of the target airway is acquired using the video laryngoscope. Next, anatomical structure recognition is performed on the airway image to obtain an anatomical structure image of the target airway. Then, temporal feature processing is performed on the first temporal signal to generate a visual mask representing changes in carbon dioxide concentration, which is superimposed on the anatomical structure image to obtain a fused display image. Finally, based on the fused display image and combined with the video laryngoscope, the positioning information of the endotracheal tube corresponding to the positional relationship between the endotracheal tube and the target airway is output. By using a gas sensor installed at the very tip of the endotracheal tube in the direction of entry, more accurate acquisition of... The carbon dioxide concentration within the target airway, and the resulting visual mask, can more accurately represent changes in carbon dioxide concentration. These changes, in turn, more accurately characterize the positional relationship between the endotracheal tube and the target airway, thus accurately locating the endotracheal tube. By acquiring airway images of the target airway through a video laryngoscope and performing anatomical structure recognition to obtain anatomical structure images, which are then superimposed on the visual mask, compared to existing methods that rely solely on video laryngoscope visualization or visual image recognition, this method, by performing anatomical structure recognition on airway images and combining it with a carbon dioxide concentration-related visual mask for tube location, reduces reliance on the operator's subjective judgment and avoids interference from difficult visual environments such as visual field obstruction, thereby improving the accuracy of endotracheal tube location.

[0060] It should be understood that the system provided in this application corresponds to the aforementioned method. The airway catheter positioning system for assisted guidance provided in this application can realize the airway catheter positioning method for assisted guidance provided in any of the embodiments of this application.

[0061] Adaptively, embodiments of this application also provide a computer device and a computer-readable storage medium.

[0062] The computer device includes: a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor; When the processor executes the computer program, it implements a method for assisted airway catheter positioning according to this application.

[0063] The computer-readable storage medium stores multiple instructions adapted for loading by a processor to execute a method for assisted airway catheter positioning according to this application.

[0064] The above description represents some embodiments of this application, providing a further detailed explanation of the purpose, technical solution, and beneficial effects of this application. It should be understood that the above-described embodiments of this application should not be construed as limiting this application. In particular, any changes, modifications, equivalent substitutions, and variations made by those skilled in the art within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for positioning an intratracheal catheter for assisted guidance, characterized in that, An endotracheal tube-assisted guidance system is used, the system comprising a video laryngoscope, an endotracheal tube, and a gas sensor, wherein the video laryngoscope is electrically connected to the gas sensor, and the gas sensor is installed at the foremost end of the endotracheal tube in the direction of insertion; the method includes: A first time-series signal characterizing carbon dioxide concentration inside the target airway and an airway image of the target airway are acquired; wherein, the first time-series signal is acquired by the gas sensor; and the airway image is acquired by the video laryngoscope. Anatomical structure recognition is performed on the airway image to obtain the anatomical structure image of the target airway; The first time-series signal is subjected to time-series feature processing to generate a visual mask for characterizing changes in carbon dioxide concentration, and the visual mask is superimposed on the anatomical structure image to generate a fused display image; Based on the fused display image and combined with the video laryngoscope, the positioning information of the endotracheal tube, corresponding to the positional relationship between the endotracheal tube and the target airway, is output.

2. The method for positioning an endotracheal tube for assisted guidance according to claim 1, characterized in that, The step of performing anatomical structure recognition on the airway image to obtain the anatomical structure image of the target airway specifically includes: Based on a pre-trained convolutional neural network, anatomical structure recognition is performed on the airway image to obtain an anatomical structure image of the target airway.

3. The method for positioning an endotracheal tube for assisted guidance according to claim 1, characterized in that, The step of performing time-series feature processing on the first time-series signal to generate a visual mask for characterizing changes in carbon dioxide concentration specifically includes: Based on a pre-trained temporal neural network, the first temporal signal is processed to obtain the carbon dioxide concentration change characteristics of the endotracheal tube located inside the target airway, and a visual mask for characterizing the carbon dioxide concentration change is generated based on the carbon dioxide concentration change characteristics.

4. The method for positioning an endotracheal tube for assisted guidance according to claim 1, characterized in that, Based on the fused display image and combined with the video laryngoscope, the method outputs the positioning information of the endotracheal tube corresponding to the positional relationship between the endotracheal tube and the target airway, specifically including: Based on a pre-trained localization neural network, the fused display image is identified to obtain the position guidance information of the endotracheal tube; Based on the fused display image and the visual mask, the positional relationship between the endotracheal tube and the target airway is determined. Using the video laryngoscope, the positional relationship between the endotracheal tube and the target airway is determined by combining the positioning guidance information and the positional relationship information, and the positioning information of the endotracheal tube corresponding to the positional relationship is output.

5. A method for positioning an endotracheal tube for assisted guidance according to claim 1, characterized in that, The endotracheal tube-assisted guidance system further includes an audible and visual cue module; before acquiring the first time-series signal inside the target airway to characterize carbon dioxide concentration and the airway image of the target airway, it also includes: The gas sensor is used to acquire a second time-series signal that characterizes the carbon dioxide concentration in the current channel where the endotracheal tube is located; Based on the second timing signal and combined with the preset carbon dioxide concentration threshold, the relationship between the current channel and the target airway and the position status of the endotracheal tube are determined, and the corresponding audio-visual prompt information is output through the audio-visual prompt module.

6. A method for positioning an endotracheal tube for assisted guidance according to claim 5, characterized in that, The step involves determining the relationship between the current channel and the target airway, and the position status of the endotracheal tube, based on the second timing signal and a preset carbon dioxide concentration threshold, and outputting corresponding audio-visual prompts through the audio-visual prompt module. Specifically, this includes: If the carbon dioxide concentration corresponding to the second timing signal is lower than the preset first carbon dioxide concentration threshold, it is determined that the current channel is different from the target airway and the position of the tracheal tube is abnormal. The corresponding warning audio-visual prompt is output through the audio-visual prompt module. If the carbon dioxide concentration corresponding to the second timing signal is not lower than the first carbon dioxide concentration threshold and is lower than the preset second carbon dioxide concentration threshold, then it is determined that the current channel is the same as the target airway and the position of the tracheal tube is abnormal. The corresponding abnormal audio-visual prompt is output through the audio-visual prompt module. If the carbon dioxide concentration corresponding to the second timing signal is not lower than the second carbon dioxide concentration threshold, then it is determined that the current channel is the same as the target airway and the position of the tracheal tube is normal. The corresponding normal audio-visual prompt is output through the audio-visual prompt module, and the second timing signal is used as the first timing signal.

7. A endotracheal tube positioning system for assisted guidance, characterized in that, An endotracheal tube-assisted guidance system is used, comprising a video laryngoscope, an endotracheal tube, and a gas sensor. The video laryngoscope is electrically connected to the gas sensor, which is installed at the foremost end of the endotracheal tube in the direction of entry. The endotracheal tube positioning system comprises a data acquisition module, a first processing module, a second processing module, and a tube positioning module. The data acquisition module is used to acquire a first time-series signal representing carbon dioxide concentration inside the target airway and an airway image of the target airway; wherein, the first time-series signal is acquired by the gas sensor; and the airway image is acquired by the video laryngoscope. The first processing module is used to perform anatomical structure recognition on the airway image to obtain an anatomical structure image of the target airway; The second processing module is used to perform time-series feature processing on the first time-series signal, generate a visual mask for characterizing changes in carbon dioxide concentration, and superimpose the visual mask onto the anatomical structure image to generate a fused display image; The endotracheal tube positioning module is used to output positioning information of the endotracheal tube corresponding to the positional relationship between the endotracheal tube and the target airway, based on the fused display image and in conjunction with the visual laryngoscope.

8. A endotracheal tube positioning system for assisted guidance according to claim 7, characterized in that, The catheter positioning module includes a first information determination unit, a second information determination unit, and a positioning information output unit; The first information determining unit is used to identify the fused display image based on a pre-trained positioning neural network to obtain the position guidance information of the endotracheal tube; The second information determining unit is used to determine the positional relationship information between the endotracheal tube and the target airway based on the fused display image and in conjunction with the visual mask; The positioning information output unit is used to output the positioning information of the endotracheal tube corresponding to the positional relationship between the endotracheal tube and the target airway, determined by the video laryngoscope in combination with the positioning guidance information and the positional relationship information.

9. A endotracheal tube positioning system for assisted guidance according to claim 7, characterized in that, The endotracheal tube-assisted guidance system also includes an audible and visual prompting module; before acquiring the time-series signal of carbon dioxide concentration inside the target airway and the airway image of the target airway, it also includes: The gas sensor is used to acquire a second time-series signal that characterizes the carbon dioxide concentration in the current channel where the endotracheal tube is located; Based on the second timing signal and combined with the preset carbon dioxide concentration threshold, the relationship between the current channel and the target airway and the position status of the endotracheal tube are determined, and the corresponding audio-visual prompt information is output through the audio-visual prompt module.

10. A endotracheal tube positioning system for assisted guidance according to claim 9, characterized in that, The step involves determining the relationship between the current channel and the target airway, and the position status of the endotracheal tube, based on the second timing signal and a preset carbon dioxide concentration threshold, and outputting corresponding audio-visual prompts through the audio-visual prompt module. Specifically, this includes: If the carbon dioxide concentration corresponding to the second timing signal is lower than the preset first carbon dioxide concentration threshold, it is determined that the current channel is different from the target airway and the position of the tracheal tube is abnormal. The corresponding warning audio-visual prompt is output through the audio-visual prompt module. If the carbon dioxide concentration corresponding to the second timing signal is not lower than the first carbon dioxide concentration threshold and is lower than the preset second carbon dioxide concentration threshold, then it is determined that the current channel is the same as the target airway and the position of the tracheal tube is abnormal. The corresponding abnormal audio-visual prompt is output through the audio-visual prompt module. If the carbon dioxide concentration corresponding to the second timing signal is not lower than the second carbon dioxide concentration threshold, then it is determined that the current channel is the same as the target airway and the position of the tracheal tube is normal. The corresponding normal audio-visual prompt is output through the audio-visual prompt module, and the second timing signal is used as the first timing signal.

Images