Image enhancement method for anesthesia puncture

Abstract

The invention discloses an image enhancement method for anesthesia puncture, and relates to the technical field of medical image processing, and the method comprises the steps: collecting a preoperative structure image of a lumbar vertebra region of a patient, combining a candidate puncture window region area to calculate a puncture window stability index, and screening out a first candidate patient image; and carrying out enhancement processing on the candidate image, automatically marking a key structure by utilizing the trained anatomical semantic recognition model, calculating a structure recognition accuracy coefficient, comparing the structure recognition accuracy coefficient with a threshold value, judging an image enhancement effect, and adjusting a strategy. And acquiring an echo image and an operation track of the puncture needle in real time, calculating an operation consistency coefficient, evaluating the stability of an operation path, and guiding operation optimization. And 4, analyzing path deviation based on a sliding window, judging the alignment condition of a puncture path and a preoperative recommended path, ensuring that the path is stable and meets the lowest monitoring requirement, and finally outputting a structure enhancement image and puncture path marking information as reference for intraoperative guidance and postoperative review.

Description

Technical Field

[0001] This invention relates to the field of medical image processing technology, specifically to an image enhancement method for anesthesia puncture. Background Technology

[0002] In clinical epidural anesthesia, spinal analgesia, and related puncture procedures, physicians need to accurately guide the puncture needle to the target anatomical region, such as the anterior border of the ligamentum flavum or the outer edge of the dural sac. To improve the accuracy and safety of punctures, ultrasound images, CT or MRI images are usually used for preoperative or intraoperative guidance of the lumbar spine anatomy. However, existing guidance methods still have significant limitations in terms of image quality, tissue layer resolution, and real-time visualization of the path.

[0003] On the one hand, imaging of the lumbar spine region is easily affected by bone obstruction, mixed soft tissue echoes, and individual patient differences, making it difficult to clearly display key anatomical structures such as the interspinous space and ligamentum flavum, and resulting in unstable puncture window positioning. On the other hand, the puncture procedure itself is a human action, which is greatly influenced by the doctor's experience and is prone to path deviation or accidental puncture, especially in cases of multiple needle insertions, anatomical variations, or changes in intraoperative posture, where conventional image enhancement methods cannot dynamically adapt to changes in the scene.

[0004] Most existing image enhancement or assisted guidance methods are limited to single-point image optimization and lack a closed-loop mechanism that integrates preoperative structural modeling, intraoperative image enhancement, and operational process monitoring and feedback adjustment. Especially during puncture, the system struggles to dynamically compare the actual operative path with the recommended path and determine deviation convergence, thus limiting the practicality of image-guided systems in high-precision surgery. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides an image enhancement method for anesthesia puncture, thereby resolving the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: an image enhancement method for anesthesia puncture, characterized by comprising the following steps:

[0007] Step 1: Acquire preoperative structural images of the patient's lumbar spine region, including ultrasound images, magnetic resonance imaging, and computed tomography images; obtain the area AC of the candidate puncture window region, calculate the puncture window stability index WSI, and compare it with the puncture window stability threshold Wth to determine whether the current patient image has a valid puncture window. If it does, it constitutes the first candidate patient image; otherwise, a strategy is generated.

[0008] Step 2: Based on the first candidate patient image, perform image enhancement, extract multidimensional semantic features of key anatomical structures of the lumbar spine, and input them into the pre-trained anatomical semantic recognition model; apply the anatomical semantic recognition model to automatically annotate the key structures in the image; count the number of successfully annotated structures Ny, calculate and obtain the structure recognition accuracy coefficient SRA, and compare and analyze it with the structure accuracy threshold Sth to determine whether the image enhancement is qualified. If it is not qualified, a strategy is given.

[0009] Step 3: During the puncture operation, acquire the puncture needle echo image and operation trajectory image at each moment during the puncture process, and record the entire operation process at a constant frame rate; obtain the spatial deviation ΔPC and the total number of operation frames Tz, calculate the operation consistency coefficient OCI, and compare it with the operation stability threshold Oth to determine whether the operation path is stable. If it is unstable, a strategy is given.

[0010] Step 4: Perform a second puncture; collect the spatial deviation ΔPC2 and the total number of frames Tz2; using a sliding window method, extract the spatial deviation value. If the value remains below the preset path deviation threshold of 0.1mm within the sliding window, it is determined that the current puncture path has stably aligned with the preoperative recommended path; simultaneously, if the number of frames Tz2 reaches the preset minimum monitoring requirement Tmin of 100 frames, the path record is considered statistically valid; select the image sequence and trajectory data corresponding to the current operation as the final result, and output the structure-enhanced image and the updated puncture path annotation information; the generated images and data serve as a formal intraoperative guidance reference document to assist the doctor's operation, and are simultaneously archived for subsequent puncture records and postoperative review.

[0011] Preferably, step one includes:

[0012] S11. Acquire preoperative structural images of the patient's lumbar spine region, including ultrasound images, magnetic resonance imaging, and computed tomography images.

[0013] S12. The preoperative structural images were registered and processed. Spatial consistency correction was performed on images of different modalities using a spatial registration algorithm to establish a fused image. Based on the fused image, a three-dimensional anatomical model of the lumbar spine was established, and tissue layering information was extracted, including the spinous process, lamina, vertebral body, ligamentum flavum, and myelin sheath region. The total intervertebral disc volume JV, intervertebral disc width JK, and ligamentum flavum thickness RH were obtained.

[0014] S13. Use image segmentation technology to distinguish continuous tissue, fracture and void regions in the ligamentum flavum and spinal sheath regions; based on three-dimensional structural connectivity analysis, use analysis algorithms to obtain the continuous tissue volume LV, and combine it with the total intervertebral space volume JV. After dimensionless processing, obtain the continuous tissue integrity coefficient CI.

[0015] S14. Identify the anatomical regions suitable for puncture in the three-dimensional anatomical model of the lumbar spine and obtain the total area ZA of the anatomical regions; combine the intervertebral space width, ligamentum flavum thickness and continuous tissue volume fraction to extract candidate puncture window regions, and calculate the area AC of the candidate puncture window regions after dimensionless processing.

[0016] Preferably, step one also includes:

[0017] S15. Based on the standard puncture location or the puncture starting point set by the doctor, locate the center of the epidural space using the three-dimensional anatomical model of the lumbar spine, mark the puncture endpoint, and obtain the puncture path vector.

[0018] S16. In the three-dimensional anatomical model of the lumbar spine, extract the spatial coordinates of the center points of the region (L1, L2, L3, L4, L5, S1) between the first lumbar vertebra L1 and the first sacral vertebra S1, and perform three-dimensional curve fitting on the coordinate points to obtain the longitudinal axis vector of the spine.

[0019] S17. Using the obtained puncture path vector and the longitudinal vector of the spine After dimensionless processing, the puncture angle Jθ is calculated and obtained.

[0020] Preferably, step one also includes:

[0021] S18. After dimensionless processing of the total area ZA of the anatomical region, the area AC of the candidate puncture window region, and the puncture angle Jθ, the puncture window stability index WSI is calculated.

[0022] S19. By setting the preset puncture window stability threshold Wth to 0.6, the puncture window stability index WSI is compared and analyzed with the puncture window stability threshold Wth to obtain the first evaluation results, including:

[0023] When the puncture window stability index WSI is greater than or equal to the puncture window stability threshold Wth, it indicates that the current patient image has an effective puncture window, constitutes the first candidate patient image, and is continuously monitored.

[0024] When the puncture window stability index (WSI) is less than the puncture window stability threshold (Wth), it indicates that the current patient image does not have an effective puncture window, posing a risk of puncture path instability. This triggers the first warning instruction and generates the first strategy: adjusting the patient's position, including increasing the flexion angle and prone posture, increasing the visibility of the intervertebral space, and re-acquiring preoperative images; adjusting the image enhancement template parameters: expanding the recommended puncture window range by 20% and increasing local contrast by 10%; and repeating the judgment and recalculation until the puncture window stability index (WSI) is less than the puncture window stability threshold (Wth).

[0025] Preferably, step two includes:

[0026] S21. Based on the first candidate patient image, perform image enhancement, including edge sharpening, contrast enhancement, and noise suppression;

[0027] S22. Based on the enhanced image, extract multidimensional semantic features of key anatomical structures of the lumbar spine, including boundaries, shape, texture and spatial position relationships, and input them into the trained anatomical semantic recognition model;

[0028] S23. Apply the anatomical semantic recognition model to automatically annotate key structures in the image, including: vertebral body boundaries, spinous processes, ligamentum flavum, and dural sac; count the number of successfully annotated structures Ny, and output the structural annotation diagram.

[0029] Preferably, step two also includes:

[0030] S24. Based on the number of successfully labeled structures Ny, combined with the theoretical number of labeled structures NL, after dimensionless processing, the structure recognition accuracy coefficient SRA is calculated and obtained.

[0031] S25. A preset structural accuracy threshold Sth is set to 0.85, and the structural recognition accuracy coefficient SRA is compared and analyzed with the structural accuracy threshold Sth to obtain the second evaluation result, including:

[0032] When the structure recognition accuracy coefficient SRA is greater than or equal to the structure accuracy threshold Sth, it indicates that the image enhancement is qualified and continuous monitoring is required.

[0033] When the structure recognition accuracy coefficient SRA < the structure accuracy threshold Sth, it indicates that the image enhancement is unqualified, and there is a risk of structural omission or recognition error. This triggers a second warning instruction and generates a second strategy: adjust the image feature extraction parameters: optimize the receptive field of the segmentation network and enhance the edge detection sensitivity; adjust the image enhancement module parameters: increase the edge sharpening weight by 15% and enhance the contrast of the structure region by 10%; re-execute the image enhancement and structure recognition operations, and repeatedly judge and recalculate until the structure recognition accuracy coefficient SRA ≥ the structure accuracy threshold Sth.

[0034] Preferably, step three includes:

[0035] S31. During the puncture operation, by integrating a puncture needle position tracking and imaging device, the puncture needle echo image and operation trajectory image at each moment during the puncture process are collected, and the entire operation process is recorded at a constant frame rate.

[0036] S32. Using displacement estimation techniques based on image registration and optical flow tracking, the needle tip position coordinates of each frame are extracted from the echo image. The actual path is compared with the preoperative suggested path through image registration. The spatial offset of the needle tip is calculated frame by frame to obtain the spatial deviation ΔPC. At the same time, the total number of frames is recorded in units of image frames to obtain the total number of operation frames Tz.

[0037] Preferably, step three also includes:

[0038] S33. After dimensionless processing using spatial deviation ΔPC and total number of operation frames Tz, the operation consistency coefficient OCI is calculated and obtained.

[0039] S34. Preset the operational stability threshold Oth to 0.7, and compare the operational consistency coefficient OCI with the operational stability threshold Oth to obtain the third evaluation results, including:

[0040] When the operational consistency coefficient OCI is greater than or equal to the operational stability threshold Oth, it indicates that the operation path is stable, the puncture is completed, and continuous monitoring is required.

[0041] When the operational consistency coefficient OCI is less than the operational stability threshold Oth, it indicates that the operational path is unstable and there is a risk of puncture deviation, triggering the third warning instruction and generating the third strategy: re-execute the image enhancement and anatomical structure annotation process; start the image enhancement adjustment instruction to automatically adjust the parameters of the image enhancement module: increase the edge sharpness by 15%, enhance the image contrast by 10%, and locally improve the noise suppression sensitivity; obtain updated enhanced images for doctors to refer to and optimize the operation.

[0042] Preferably, step four includes:

[0043] S41. Based on the updated structural enhancement image, automatically re-annotate the optimal puncture window region and guide path;

[0044] S42. Based on the updated enhanced images and annotations, the doctor performs a second puncture; at the same time, the doctor continuously acquires the needle insertion resistance signal, echo image sequence, and operation trajectory image.

[0045] S43. Using displacement estimation techniques based on image registration and optical flow tracking, extract the needle tip position of each frame from the echo image of the secondary operation, calculate the spatial deviation ΔPC2 between the actual needle tip path and the updated guide path frame by frame, and simultaneously count the total number of operation frames Tz2.

[0046] Preferably, step four also includes:

[0047] S44. The spatial deviation ΔPC2 value of multiple consecutive frames during the entire puncture operation is extracted using a sliding window method, and the fluctuation range of the spatial deviation ΔPC2 value in multiple consecutive frames is statistically analyzed. When the spatial deviation value gradually stabilizes within the sliding window and remains below the preset path deviation threshold of 0.1mm, the system automatically determines that the current puncture path has been stably aligned with the preoperative recommended path. At the same time, if the number of frames Tz2 reaches the preset minimum monitoring requirement Tmin of 100 frames, the path record is considered to have statistical validity.

[0048] S45. When the puncture path deviation converges and stabilizes, and the number of frames acquired meets the monitoring requirements, the system selects the image sequence and trajectory data corresponding to the current operation as the final result, and outputs the structure-enhanced image and the updated puncture path annotation information; the generated images and data serve as formal intraoperative guidance reference documents to assist doctors in their operations, and are simultaneously archived for subsequent puncture records and postoperative review.

[0049] This invention provides an image enhancement method for anesthesia puncture. It has the following beneficial effects:

[0050] (1) This image enhancement method for anesthesia puncture introduces quantitative indicators such as puncture window stability index (WSI), structural recognition accuracy coefficient (SRA), and operation consistency coefficient (OCI), and combines image registration and three-dimensional modeling technology to achieve quantitative analysis and evaluation of the patient's preoperative images. It can effectively identify whether there is a puncture window, significantly improve the usability and evaluation accuracy of preoperative images, and assist doctors in personalized path planning and position adjustment.

[0051] (2) The image enhancement method for anesthesia puncture automatically extracts and labels key structures such as spinous process, ligamentum flavum, and dural sac through image enhancement processing and anatomical semantic recognition model to form structure enhancement images, realize intelligent visual guidance for puncture operation, greatly improve doctors' ability to recognize complex structural areas, reduce the risk of accidental puncture, and improve the safety of puncture operation.

[0052] (3) The image enhancement method for anesthesia puncture combines image acquisition, spatial deviation calculation and operation consistency evaluation mechanism during puncture to construct a dynamic closed-loop control process; when the operation path deviates or the image recognition is unstable, the system can automatically issue an early warning and execute enhancement parameter adjustment and image reconstruction strategy to realize the adaptive adjustment of the image enhancement module and ensure that the operation process is always in a stable guiding state.

[0053] (4) This image enhancement method for anesthesia puncture, through the acquisition of images and trajectories of the entire process of secondary puncture operation, after the spatial deviation is stable and converged and the number of frames meets the statistical requirements, selects the current image and path data as the formal result output, automatically generates intraoperative guidance documents, realizes the standardized archiving and traceability of image enhancement guidance results, and helps postoperative quality assessment and puncture technique improvement. Attached Figure Description

[0054] Figure 1 This is a schematic diagram of the steps of an image enhancement method for anesthesia puncture according to the present invention. Detailed Implementation

[0055] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0056] Example 1

[0057] Please see Figure 1 The present invention provides an image enhancement method for anesthesia puncture, comprising the following steps:

[0058] Step 1: Acquire preoperative structural images of the patient's lumbar spine region, including ultrasound images, magnetic resonance imaging, and computed tomography images; obtain the area AC of the candidate puncture window region, calculate the puncture window stability index WSI, and compare it with the puncture window stability threshold Wth to determine whether the current patient image has a valid puncture window. If it does, it constitutes the first candidate patient image; otherwise, a strategy is generated.

[0059] Step 2: Based on the first candidate patient image, perform image enhancement, extract multidimensional semantic features of key anatomical structures of the lumbar spine, and input them into the pre-trained anatomical semantic recognition model; apply the anatomical semantic recognition model to automatically annotate the key structures in the image; count the number of successfully annotated structures Ny, calculate and obtain the structure recognition accuracy coefficient SRA, and compare and analyze it with the structure accuracy threshold Sth to determine whether the image enhancement is qualified. If it is not qualified, a strategy is given.

[0060] Step 3: During the puncture operation, acquire the puncture needle echo image and operation trajectory image at each moment during the puncture process, and record the entire operation process at a constant frame rate; obtain the spatial deviation ΔPC and the total number of operation frames Tz, calculate the operation consistency coefficient OCI, and compare it with the operation stability threshold Oth to determine whether the operation path is stable. If it is unstable, a strategy is given.

[0061] Step 4: Perform a second puncture; collect the spatial deviation ΔPC2 and the total number of frames Tz2; using a sliding window method, extract the spatial deviation value. If the value remains below the preset path deviation threshold of 0.1mm within the sliding window, it is determined that the current puncture path has stably aligned with the preoperative recommended path; simultaneously, if the number of frames Tz2 reaches the preset minimum monitoring requirement Tmin of 100 frames, the path record is considered statistically valid; select the image sequence and trajectory data corresponding to the current operation as the final result, and output the structure-enhanced image and the updated puncture path annotation information; the generated images and data serve as a formal intraoperative guidance reference document to assist the doctor's operation, and are simultaneously archived for subsequent puncture records and postoperative review.

[0062] In this embodiment, by calculating the puncture window stability index (WSI) and setting a stability threshold (Wth), a quantitative assessment of the candidate puncture area in the patient's lumbar spine image was achieved before the operation. This allows for the identification of whether a stable and reliable puncture window exists at the initial stage of image acquisition, effectively reducing the risk of puncture failure due to differences in anatomical structure, and providing a scientific basis for subsequent image enhancement and path planning.

[0063] Example 2

[0064] This embodiment is an explanation based on Embodiment 1. Please refer to it. Figure 1 Specifically, step one includes:

[0065] S11. Acquire preoperative structural images of the patient's lumbar spine region, including ultrasound images, magnetic resonance imaging, and computed tomography images.

[0066] S12. The preoperative structural images were registered and processed. Spatial consistency correction was performed on images of different modalities using a spatial registration algorithm to establish a fused image. Based on the fused image, a three-dimensional anatomical model of the lumbar spine was established, and tissue layering information was extracted, including the spinous process, lamina, vertebral body, ligamentum flavum, and myelin sheath region. The total intervertebral disc volume JV, intervertebral disc width JK, and ligamentum flavum thickness RH were obtained.

[0067] S13. Image segmentation techniques are used to distinguish continuous tissue, fractured, and void regions in the ligamentum flavum and spinal sheath regions. Based on three-dimensional structural connectivity analysis, an analysis algorithm is used to obtain the continuous tissue volume LV, and combined with the total intervertebral space volume JV, after dimensionless processing, the continuous tissue integrity coefficient CI is obtained, as shown in the following formula:

[0068]

[0069] S14. Identify the anatomical regions suitable for puncture in the three-dimensional anatomical model of the lumbar spine and obtain the total area ZA of the anatomical regions; combine the intervertebral space width, ligamentum flavum thickness, and continuous tissue volume fraction to extract candidate puncture window regions. After dimensionless processing, calculate the area AC of the candidate puncture window regions using the following formula:

[0070]

[0071] In the formula, N represents the total number of candidate intervertebral spaces, and JK i RH represents the width of the i-th intervertebral space. i CI represents the thickness of the ligamentum flavum in the i-th intervertebral space. i This represents the continuity coefficient of tissue integrity in the i-th intervertebral space channel.

[0072] In this embodiment, a precise three-dimensional anatomical model of the lumbar spine is constructed through spatial registration and fusion of multimodal images. By combining layered tissue information and continuous tissue integrity coefficient, the intervertebral space and puncture-related anatomical features are scientifically quantified, enabling accurate identification and area calculation of candidate puncture window regions. This significantly improves the accuracy and reliability of puncture window assessment, providing a solid imaging foundation for subsequent puncture path planning and surgical safety.

[0073] Example 3

[0074] This embodiment is an explanation based on Embodiment 2. Please refer to it. Figure 1 Specifically, step one also includes:

[0075] S15. Based on the standard puncture location or the puncture starting point set by the doctor, locate the center of the epidural space using the three-dimensional anatomical model of the lumbar spine, mark the puncture endpoint, and obtain the puncture path vector.

[0076] S16. In the three-dimensional anatomical model of the lumbar spine, extract the spatial coordinates of the center points of the region (L1, L2, L3, L4, L5, S1) between the first lumbar vertebra L1 and the first sacral vertebra S1, and perform three-dimensional curve fitting on the coordinate points to obtain the longitudinal axis vector of the spine.

[0077] S17. Using the obtained puncture path vector and the longitudinal vector of the spine After dimensionless processing, the puncture angle Jθ is calculated using the following formula:

[0078]

[0079] In this embodiment, based on a precise three-dimensional anatomical model of the lumbar spine, combined with the standard puncture starting point and epidural cavity center positioning, the longitudinal axis vector of the spine is accurately extracted using a three-dimensional curve fitting method. By calculating the angle between the puncture path vector and the longitudinal axis vector of the spine, the precise spatial orientation and optimization of the puncture path are achieved, effectively improving the accuracy and safety of the puncture operation and reducing the risk of puncture deviation and complications.

[0080] Example 5

[0081] This embodiment is an explanation based on Embodiment 1. Please refer to it. Figure 1 Specifically, step one also includes:

[0082] S18. The puncture window stability index (WSI) is calculated after dimensionless processing using the total area ZA of the anatomical region, the area AC of the candidate puncture window region, and the puncture angle Jθ. The formula is as follows:

[0083]

[0084] S19. By setting the preset puncture window stability threshold Wth to 0.6, the puncture window stability index WSI is compared and analyzed with the puncture window stability threshold Wth to obtain the first evaluation results, including:

[0085] When the puncture window stability index WSI is greater than or equal to the puncture window stability threshold Wth, it indicates that the current patient image has an effective puncture window, constitutes the first candidate patient image, and is continuously monitored.

[0086] When the puncture window stability index (WSI) is less than the puncture window stability threshold (Wth), it indicates that the current patient image does not have an effective puncture window, posing a risk of puncture path instability. This triggers the first warning instruction and generates the first strategy: adjusting the patient's position, including increasing the flexion angle and prone posture, increasing the visibility of the intervertebral space, and re-acquiring preoperative images; adjusting the image enhancement template parameters: expanding the recommended puncture window range by 20% and increasing local contrast by 10%; and repeating the judgment and recalculation until the puncture window stability index (WSI) is less than the puncture window stability threshold (Wth).

[0087] The method for setting the puncture window stability threshold Wth is based on statistical analysis of preoperative imaging data and surgical success rates from a large number of lumbar puncture surgeries, comprehensively considering the puncture difficulty and success probability corresponding to different puncture window stability indices (WSI). Combining orthopedic surgical imaging research findings, clinical puncture experience, and relevant anatomical standards, a scientifically reasonable threshold classification was established. Through multi-center clinical data validation, Wth of 0.6 was determined as the key critical value for distinguishing between effective and ineffective puncture windows, guiding preoperative assessment and puncture path planning, and improving puncture safety and accuracy.

[0088] In this embodiment, the effectiveness of the puncture window is accurately assessed by constructing a puncture window stability index (WSI) and dynamically comparing it with a preset threshold (Wth). When the index is lower than the threshold, a closed-loop optimization strategy is automatically triggered to adjust the patient's position and image enhancement parameters, which significantly improves the stability and visualization effect of the puncture window, ensures the safety of the puncture path and the success rate of the operation, and effectively reduces the risk of puncture failure.

[0089] Example 5

[0090] This embodiment is an explanation based on Embodiment 4. Please refer to it. Figure 1 Specifically, step two includes:

[0091] S21. Based on the first candidate patient image, perform image enhancement, including edge sharpening, contrast enhancement, and noise suppression;

[0092] S22. Based on the enhanced image, extract multidimensional semantic features of key anatomical structures of the lumbar spine, including boundaries, shape, texture and spatial position relationships, and input them into the trained anatomical semantic recognition model;

[0093] S23. Apply the anatomical semantic recognition model to automatically annotate key structures in the image, including: vertebral body boundaries, spinous processes, ligamentum flavum, and dural sac; count the number of successfully annotated structures Ny, and output the structural annotation diagram.

[0094] In this embodiment, by performing multidimensional image enhancement on the first candidate patient image, the edge clarity and contrast of key anatomical structures of the lumbar spine are effectively improved. Combined with a well-trained anatomical semantic recognition model, the automatic and accurate annotation of key structures such as vertebral body boundaries, spinous processes, ligamentum flavum, and dural sac is achieved, which significantly improves the accuracy and efficiency of structure recognition and provides reliable image support for subsequent puncture path planning and surgical navigation.

[0095] Example 6

[0096] This embodiment is an explanation based on Embodiment 5. Please refer to it. Figure 1 Specifically, step two also includes:

[0097] S24. Based on the number of successfully labeled structures Ny, combined with the theoretically required number of labeled structures NL from the model, and after dimensionless processing, the structure recognition accuracy coefficient SRA is calculated as follows:

[0098]

[0099] S25. A preset structural accuracy threshold Sth is set to 0.85, and the structural recognition accuracy coefficient SRA is compared and analyzed with the structural accuracy threshold Sth to obtain the second evaluation result, including:

[0100] When the structure recognition accuracy coefficient SRA is greater than or equal to the structure accuracy threshold Sth, it indicates that the image enhancement is qualified and continuous monitoring is required.

[0101] When the structure recognition accuracy coefficient SRA < the structure accuracy threshold Sth, it indicates that the image enhancement is unqualified, and there is a risk of structural omission or recognition error. This triggers a second warning instruction and generates a second strategy: adjust the image feature extraction parameters: optimize the receptive field of the segmentation network and enhance the edge detection sensitivity; adjust the image enhancement module parameters: increase the edge sharpening weight by 15% and enhance the contrast of the structure region by 10%; re-execute the image enhancement and structure recognition operations, and repeatedly judge and recalculate until the structure recognition accuracy coefficient SRA ≥ the structure accuracy threshold Sth.

[0102] The method for setting the structural accuracy threshold Sth is as follows: It is derived from the statistical results of model training and validation on a large number of lumbar spine structural image enhancements and automatic anatomical structure recognition, combined with the analysis of the impact of different recognition accuracy rates on puncture assistance effects. Referring to industry standards in the field of medical image segmentation and recognition, model performance evaluation indicators, and the actual needs of clinicians for annotation accuracy, a reasonable accuracy threshold was comprehensively determined. Based on clinical trial feedback and expert review, Sth was ultimately set to 0.85 to ensure the reliability and practicality of automatic anatomical structure annotation, thereby supporting effective intraoperative navigation and puncture guidance.

[0103] In this embodiment, the structure recognition accuracy coefficient (SRA) is calculated and dynamically compared with a preset threshold (Sth) to achieve real-time evaluation of the image enhancement effect, effectively detecting the risk of structure omissions and recognition errors. For cases of unqualified recognition, an optimization strategy is automatically triggered, adjusting image feature extraction and enhancement parameters, iteratively improving the accuracy and completeness of structure annotation, ensuring precise identification of key anatomical structures, and providing more reliable image support and assurance for puncture operations.

[0104] Example 7

[0105] This embodiment is an explanation based on Embodiment 6. Please refer to it. Figure 1 Specifically, step three includes:

[0106] S31. During the puncture operation, by integrating a puncture needle position tracking and imaging device, the puncture needle echo image and operation trajectory image at each moment during the puncture process are collected, and the entire operation process is recorded at a constant frame rate.

[0107] S32. Using displacement estimation techniques based on image registration and optical flow tracking, the needle tip position coordinates of each frame are extracted from the echo image. The actual path is compared with the preoperative suggested path through image registration. The spatial offset of the needle tip is calculated frame by frame to obtain the spatial deviation ΔPC. At the same time, the total number of frames is recorded in units of image frames to obtain the total number of operation frames Tz.

[0108] In this embodiment, by integrating a puncture needle pose tracking and imaging device, real-time echo images and operation trajectory images of the puncture needle are acquired. Combined with image registration and optical flow tracking technology, the needle tip position coordinates are accurately extracted, enabling frame-by-frame spatial deviation calculation of the puncture path and high-frame-rate recording of the entire operation. This technology effectively improves the accuracy and timeliness of puncture path monitoring, provides a scientific basis for dynamic adjustments during surgery, and significantly enhances the safety and stability of the puncture process.

[0109] Example 8

[0110] This embodiment is an explanation based on Embodiment 7. Please refer to it. Figure 1 Specifically, step three also includes:

[0111] S33. After dimensionless processing using the spatial deviation ΔPC and the total number of operation frames Tz, the operation consistency coefficient OCI is calculated as follows:

[0112]

[0113] S34. Preset the operational stability threshold Oth to 0.7, and compare the operational consistency coefficient OCI with the operational stability threshold Oth to obtain the third evaluation results, including:

[0114] When the operational consistency coefficient OCI is greater than or equal to the operational stability threshold Oth, it indicates that the operation path is stable, the puncture is completed, and continuous monitoring is required.

[0115] When the operational consistency coefficient OCI is less than the operational stability threshold Oth, it indicates that the operational path is unstable and there is a risk of puncture deviation, triggering the third warning instruction and generating the third strategy: re-execute the image enhancement and anatomical structure annotation process; start the image enhancement adjustment instruction to automatically adjust the parameters of the image enhancement module: increase the edge sharpness by 15%, enhance the image contrast by 10%, and locally improve the noise suppression sensitivity; obtain updated enhanced images for doctors to refer to and optimize the operation.

[0116] The method for setting the operational stability threshold Oth is based on a correlation study between the consistency of the puncture path during the procedure and the surgical success rate, combined with echocardiographic data and real-time monitoring data of the puncture trajectory. Statistical analysis of multiple surgical trajectory data was conducted to evaluate the relationship between the operational consistency coefficient (OCI) and puncture accuracy. Referring to surgical procedure standards and intraoperative monitoring technical specifications, and combining expert experience, Oth was determined to be 0.7 as the criterion for judging operational path stability, to guide real-time intraoperative feedback and operational optimization, and improve the safety and stability of puncture procedures.

[0117] In this embodiment, the stability of the puncture path is dynamically assessed by calculating the Operation Consistency Coefficient (OCI) and comparing it with a preset Operation Stability Threshold (Oth). When the operation path is unstable, an early warning is automatically triggered, and an image enhancement adjustment strategy is activated to improve edge sharpening, image contrast, and noise suppression, thereby optimizing the visual representation of anatomical structures. This assists the physician in adjusting the operation path, significantly reducing the risk of puncture deviation and improving the safety and success rate of the puncture procedure.

[0118] Example 9

[0119] This embodiment is an explanation based on Embodiment 8. Please refer to it. Figure 1 Specifically, step four includes:

[0120] S41. Based on the updated structural enhancement image, automatically re-annotate the optimal puncture window region and guide path;

[0121] S42. Based on the updated enhanced images and annotations, the doctor performs a second puncture; at the same time, the doctor continuously acquires the needle insertion resistance signal, echo image sequence, and operation trajectory image.

[0122] S43. Using displacement estimation techniques based on image registration and optical flow tracking, extract the needle tip position of each frame from the echo image of the secondary operation, calculate the spatial deviation ΔPC2 between the actual needle tip path and the updated guide path frame by frame, and simultaneously count the total number of operation frames Tz2.

[0123] In this embodiment, the optimal puncture window area and guide path are automatically re-marked based on the updated structure-enhanced image. Combined with the real-time acquired puncture needle resistance signal, echo image sequence and operation trajectory, the system can accurately monitor the secondary puncture process and dynamically calculate the path deviation, effectively improving the accuracy and safety of the puncture operation. This assists doctors in adjusting their operation strategies in a timely manner, ensuring that the puncture path is highly consistent with the preoperative plan and reducing surgical risks.

[0124] Example 10

[0125] This embodiment is an explanation based on Embodiment 9. Please refer to it. Figure 1 Specifically, step four also includes:

[0126] S44. The spatial deviation ΔPC2 value of multiple consecutive frames during the entire puncture operation is extracted using a sliding window method, and the fluctuation range of the spatial deviation ΔPC2 value in multiple consecutive frames is statistically analyzed. When the spatial deviation value gradually stabilizes within the sliding window and remains below the preset path deviation threshold of 0.1mm, the system automatically determines that the current puncture path has been stably aligned with the preoperative recommended path. At the same time, if the number of frames Tz2 reaches the preset minimum monitoring requirement Tmin of 100 frames, the path record is considered to have statistical validity.

[0127] S45. When the puncture path deviation converges and stabilizes, and the number of frames acquired meets the monitoring requirements, the system selects the image sequence and trajectory data corresponding to the current operation as the final result, and outputs the structure-enhanced image and the updated puncture path annotation information; the generated images and data serve as formal intraoperative guidance reference documents to assist doctors in their operations, and are simultaneously archived for subsequent puncture records and postoperative review.

[0128] The path deviation threshold was set based on research into the impact of spatial deviation on puncture safety and accuracy during actual needle manipulation. By analyzing the risk of nerve injury and puncture success rate within different deviation ranges, combined with image navigation accuracy and mechanical control errors, and referencing clinical minimally invasive surgery accuracy standards and international medical image navigation guidelines, strict path deviation control indicators were formulated. The deviation threshold was set below 0.1 mm to ensure high-precision alignment of the puncture path, reduce the probability of complications, and ensure patient safety.

[0129] The minimum monitoring frame count Tmin is set based on statistical analysis of image acquisition frequency and path stability during the puncture procedure. Through sliding window analysis of a large amount of puncture procedure video data, combined with statistical significance testing, it was determined that spatial deviation assessment results have sufficient stability and representativeness within a continuous data range of 100 frames or more. Referring to medical equipment image acquisition standards and intraoperative monitoring technical specifications, and considering actual operation time and physician operating habits, Tmin is set at 100 frames to ensure the accuracy and reliability of path stability determination.

[0130] In this embodiment, the spatial deviation of the puncture path is dynamically monitored by using a sliding window method. The system determines in real time whether the path deviation is stable and remains below the strict 0.1mm threshold. Combined with the minimum frame rate monitoring requirement, the statistical validity of the path alignment is ensured. Finally, a high-precision structure-enhanced image and updated puncture path annotation information are output to form a complete intraoperative guidance reference document. This not only improves the accuracy and safety of the puncture operation, but also provides reliable data support for postoperative review and quality tracking.

[0131] The threshold is set to facilitate comparison. The size of the threshold depends on the amount of sample data and the number of bases set by those skilled in the art for each set of sample data; as long as it does not affect the ratio between the parameter and the quantized value, it is acceptable.

[0132] The above formulas are all derived from software simulation using a large amount of data, and are selected to be close to the actual values. The coefficients in the formulas are set by those skilled in the art based on the actual situation. The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or changes made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. An image enhancement method for anesthesia puncture, characterized in that, Includes the following steps: Step 1: Acquire preoperative structural images of the patient's lumbar spine region, including ultrasound images, magnetic resonance imaging, and computed tomography images; obtain the area AC of the candidate puncture window region, calculate the puncture window stability index WSI, and compare it with the puncture window stability threshold Wth to determine whether the current patient image has a valid puncture window. If it does, it constitutes the first candidate patient image; otherwise, a strategy is generated. Step 2: Based on the first candidate patient image, perform image enhancement, extract multidimensional semantic features of key anatomical structures of the lumbar spine, and input them into the pre-trained anatomical semantic recognition model; apply the anatomical semantic recognition model to automatically annotate the key structures in the image; count the number of successfully annotated structures Ny, calculate and obtain the structure recognition accuracy coefficient SRA, and compare and analyze it with the structure accuracy threshold Sth to determine whether the image enhancement is qualified. If it is not qualified, a strategy is given. Step 3: During the puncture operation, acquire the puncture needle echo image and operation trajectory image at each moment during the puncture process, and record the entire operation process at a constant frame rate; obtain the spatial deviation ΔPC and the total number of operation frames Tz, calculate the operation consistency coefficient OCI, and compare it with the operation stability threshold Oth to determine whether the operation path is stable. If it is unstable, a strategy is given. Step 4: Perform a second puncture; collect the spatial deviation ΔPC2 and the total number of frames Tz2; using a sliding window method, extract the spatial deviation value. If the value remains below the preset path deviation threshold of 0.1mm within the sliding window, it is determined that the current puncture path has stably aligned with the preoperative recommended path; simultaneously, if the number of frames Tz2 reaches the preset minimum monitoring requirement Tmin of 100 frames, the path record is considered statistically valid; select the image sequence and trajectory data corresponding to the current operation as the final result, and output the structure-enhanced image and the updated puncture path annotation information; the generated images and data serve as a formal intraoperative guidance reference document to assist the doctor's operation, and are simultaneously archived for subsequent puncture records and postoperative review.

2. The image enhancement method for anesthesia puncture according to claim 1, characterized in that, Step one includes: S11. Acquire preoperative structural images of the patient's lumbar spine region, including ultrasound images, magnetic resonance imaging, and computed tomography images. S12. The preoperative structural images were registered and processed. Spatial consistency correction was performed on images of different modalities using a spatial registration algorithm to establish a fused image. Based on the fused image, a three-dimensional anatomical model of the lumbar spine was established, and tissue layering information was extracted, including the spinous process, lamina, vertebral body, ligamentum flavum, and myelin sheath region. The total intervertebral disc volume JV, intervertebral disc width JK, and ligamentum flavum thickness RH were obtained. S13. Use image segmentation technology to distinguish continuous tissue, fracture and void regions in the ligamentum flavum and spinal sheath regions; based on three-dimensional structural connectivity analysis, use analysis algorithms to obtain the continuous tissue volume LV, and combine it with the total intervertebral space volume JV. After dimensionless processing, obtain the continuous tissue integrity coefficient CI. S14. Identify the anatomical regions suitable for puncture in the three-dimensional anatomical model of the lumbar spine and obtain the total area ZA of the anatomical regions; combine the intervertebral space width, ligamentum flavum thickness and continuous tissue volume fraction to extract candidate puncture window regions, and calculate the area AC of the candidate puncture window regions after dimensionless processing.

3. The image enhancement method for anesthesia puncture according to claim 2, characterized in that, Step one also includes: S15. Based on the standard puncture location or the puncture starting point set by the doctor, locate the center of the epidural space using the three-dimensional anatomical model of the lumbar spine, mark the puncture endpoint, and obtain the puncture path vector. S16. In the three-dimensional anatomical model of the lumbar spine, extract the spatial coordinates of the center points of the region (L1, L2, L3, L4, L5, S1) between the first lumbar vertebra L1 and the first sacral vertebra S1, and perform three-dimensional curve fitting on the coordinate points to obtain the longitudinal axis vector of the spine. S17. Using the obtained puncture path vector and the longitudinal vector of the spine After dimensionless processing, the puncture angle Jθ is calculated and obtained.

4. The image enhancement method for anesthesia puncture according to claim 3, characterized in that, Step one also includes: S18. After dimensionless processing of the total area ZA of the anatomical region, the area AC of the candidate puncture window region, and the puncture angle Jθ, the puncture window stability index WSI is calculated. S19. By setting the preset puncture window stability threshold Wth to 0.6, the puncture window stability index WSI is compared and analyzed with the puncture window stability threshold Wth to obtain the first evaluation results, including: When the puncture window stability index WSI is greater than or equal to the puncture window stability threshold Wth, it indicates that the current patient image has an effective puncture window, constitutes the first candidate patient image, and is continuously monitored. When the puncture window stability index (WSI) is less than the puncture window stability threshold (Wth), it indicates that the current patient image does not have an effective puncture window, posing a risk of puncture path instability. This triggers the first warning instruction and generates the first strategy: adjusting the patient's position, including increasing the flexion angle and prone posture, increasing the visibility of the intervertebral space, and re-acquiring preoperative images; adjusting the image enhancement template parameters: expanding the recommended puncture window range by 20% and increasing local contrast by 10%; and repeating the judgment and recalculation until the puncture window stability index (WSI) is less than the puncture window stability threshold (Wth).

5. The image enhancement method for anesthesia puncture according to claim 4, characterized in that, Step two includes: S21. Based on the first candidate patient image, perform image enhancement, including edge sharpening, contrast enhancement, and noise suppression; S22. Based on the enhanced image, extract multidimensional semantic features of key anatomical structures of the lumbar spine, including boundaries, shape, texture and spatial position relationships, and input them into the trained anatomical semantic recognition model; S23. Apply the anatomical semantic recognition model to automatically annotate key structures in the image, including: vertebral body boundaries, spinous processes, ligamentum flavum, and dural sac; count the number of successfully annotated structures Ny, and output the structural annotation diagram.

6. The image enhancement method for anesthesia puncture according to claim 5, characterized in that, Step two also includes: S24. Based on the number of successfully labeled structures Ny, combined with the theoretical number of labeled structures NL, after dimensionless processing, the structure recognition accuracy coefficient SRA is calculated and obtained. S25. A preset structural accuracy threshold Sth is set to 0.85, and the structural recognition accuracy coefficient SRA is compared and analyzed with the structural accuracy threshold Sth to obtain the second evaluation result, including: When the structure recognition accuracy coefficient SRA is greater than or equal to the structure accuracy threshold Sth, it indicates that the image enhancement is qualified and continuous monitoring is required. When the structure recognition accuracy coefficient SRA < the structure accuracy threshold Sth, it indicates that the image enhancement is unqualified, and there is a risk of structural omission or recognition error. This triggers a second warning instruction and generates a second strategy: adjust the image feature extraction parameters: optimize the receptive field of the segmentation network and enhance the edge detection sensitivity; adjust the image enhancement module parameters: increase the edge sharpening weight by 15% and enhance the contrast of the structure region by 10%; re-execute the image enhancement and structure recognition operations, and repeatedly judge and recalculate until the structure recognition accuracy coefficient SRA ≥ the structure accuracy threshold Sth.

7. The image enhancement method for anesthesia puncture according to claim 6, characterized in that, Step three includes: S31. During the puncture operation, by integrating a puncture needle position tracking and imaging device, the puncture needle echo image and operation trajectory image at each moment during the puncture process are collected, and the entire operation process is recorded at a constant frame rate. S32. Using displacement estimation techniques based on image registration and optical flow tracking, the needle tip position coordinates of each frame are extracted from the echo image. The actual path is compared with the preoperative suggested path through image registration. The spatial offset of the needle tip is calculated frame by frame to obtain the spatial deviation ΔPC. At the same time, the total number of frames is recorded in units of image frames to obtain the total number of operation frames Tz.

8. The image enhancement method for anesthesia puncture according to claim 7, characterized in that, Step three also includes: S33. After dimensionless processing using spatial deviation ΔPC and total number of operation frames Tz, the operation consistency coefficient OCI is calculated and obtained. S34. Preset the operational stability threshold Oth to 0.7, and compare the operational consistency coefficient OCI with the operational stability threshold Oth to obtain the third evaluation results, including: When the operational consistency coefficient OCI is greater than or equal to the operational stability threshold Oth, it indicates that the operation path is stable, the puncture is completed, and continuous monitoring is required. When the operational consistency coefficient OCI is less than the operational stability threshold Oth, it indicates that the operational path is unstable and there is a risk of puncture deviation, triggering the third warning instruction and generating the third strategy: re-execute the image enhancement and anatomical structure annotation process; start the image enhancement adjustment instruction to automatically adjust the parameters of the image enhancement module: increase the edge sharpness by 15%, enhance the image contrast by 10%, and locally improve the noise suppression sensitivity; obtain updated enhanced images for doctors to refer to and optimize the operation.

9. The image enhancement method for anesthesia puncture according to claim 8, characterized in that, Step four includes: S41. Based on the updated structural enhancement image, automatically re-annotate the optimal puncture window region and guide path; S42. Based on the updated enhanced images and annotations, the doctor performs a second puncture; at the same time, the doctor continuously acquires the needle insertion resistance signal, echo image sequence, and operation trajectory image. S43. Using displacement estimation techniques based on image registration and optical flow tracking, extract the needle tip position of each frame from the echo image of the secondary operation, calculate the spatial deviation ΔPC2 between the actual needle tip path and the updated guide path frame by frame, and simultaneously count the total number of operation frames Tz2.

10. The image enhancement method for anesthesia puncture according to claim 9, characterized in that, Step four also includes: S44. The spatial deviation ΔPC2 value of multiple consecutive frames during the entire puncture operation is extracted using a sliding window method, and the fluctuation range of the spatial deviation ΔPC2 value in multiple consecutive frames is statistically analyzed. When the spatial deviation value gradually stabilizes within the sliding window and remains below the preset path deviation threshold of 0.1mm, the system automatically determines that the current puncture path has been stably aligned with the preoperative recommended path. At the same time, if the number of frames Tz2 reaches the preset minimum monitoring requirement Tmin of 100 frames, the path record is considered to have statistical validity. S45. When the puncture path deviation converges and stabilizes, and the number of frames acquired meets the monitoring requirements, the system selects the image sequence and trajectory data corresponding to the current operation as the final result, and outputs the structure-enhanced image and the updated puncture path annotation information; the generated images and data serve as formal intraoperative guidance reference documents to assist doctors in their operations, and are simultaneously archived for subsequent puncture records and postoperative review.

Images