PICC catheter positioning and recognizing device and method based on optical coherence tomography

Abstract

The application discloses a PICC catheterization positioning and recognition device and method based on optical coherence tomography, and the device comprises an OCT probe, a signal acquisition and processing unit, an image display unit and a position recognition unit; the OCT probe is used for acquiring OCT images of blood vessels and a PICC catheter in a PICC catheterization process; the signal acquisition and processing unit is used for processing the acquired OCT image signals to obtain processed OCT images; the image display unit is used for displaying the processed OCT images in real time; and the position recognition unit is used for analyzing and recognizing the processed OCT images to obtain real-time position information of the PICC catheter, comparing the real-time position information with a preset optimal position, and obtaining a position discrimination result.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of medical imaging, in particular to a PICC catheter positioning and recognition device and method based on optical coherence tomography. BACKGROUND

[0002] In clinical medicine, PICC (peripherally inserted central catheter) catheterization is a common operation used to provide long-term intravenous infusion access for patients. Accurate PICC catheter positioning is crucial for ensuring treatment effectiveness and patient safety. Currently, ultrasound is a commonly used method for PICC catheter positioning, but ultrasound imaging may be affected by factors such as patient size and operator experience, resulting in inaccurate positioning. Therefore, a new real-time imaging technology is needed that can provide clearer and more accurate intravascular images during catheterization, and the device should be portable to facilitate use in different medical settings. SUMMARY

[0003] The present application aims to provide a PICC catheter positioning and recognition device and method based on optical coherence tomography (OCT) to solve the problem of inaccurate PICC catheter positioning and inconvenient equipment in the prior art.

[0004] In a first aspect of the present application, a PICC catheter positioning and recognition device based on optical coherence tomography (OCT) is disclosed, comprising: an OCT probe, a signal acquisition and processing unit, an image display unit, and a position recognition unit.

[0005] The OCT probe is used to acquire OCT images of blood vessels and PICC catheters during the PICC catheterization process.

[0006] The signal acquisition and processing unit is connected to the OCT probe and is used to process the acquired OCT image signals to obtain processed OCT images.

[0007] The image display unit is connected to the signal acquisition and processing unit and is used to display the processed OCT images in real time.

[0008] The position recognition unit is connected to the image display unit and the signal acquisition and processing unit and is used to analyze and recognize the processed OCT images to obtain real-time position information of the PICC catheter, compare it with the preset optimal position, and obtain a position discrimination result.

[0009] The OCT probe is implemented using a cardiovascular OCT imaging device.

[0010] The signal acquisition and processing unit is used to process the acquired OCT image signals to obtain processed OCT images, including:

[0011] The signal acquisition processing unit performs image denoising processing on the acquired OCT image signal to obtain a first image signal;

[0012] The first image signal is subjected to contrast enhancement processing to obtain a second image signal;

[0013] The second image signal is subjected to correction processing to obtain a processed OCT image.

[0014] In a second aspect, the application discloses a PICC catheter positioning and identifying method based on optical coherence tomography, which is implemented by using the PICC catheter positioning and identifying device based on optical coherence tomography, and comprises the following steps:

[0015] S1, using an OCT probe, acquiring OCT images of blood vessels and a PICC catheter in a PICC catheterization process;

[0016] S2, using a signal acquisition processing unit, processing the acquired OCT image signal to obtain a processed OCT image;

[0017] S3, using an image display unit, displaying the processed OCT image in real time;

[0018] S4, using a position identifying unit, analyzing and identifying the processed OCT image to obtain real-time position information of the PICC catheter, comparing the real-time position information of the PICC catheter with a preset optimal position to obtain a position identification result.

[0019] The processing of the acquired OCT image signal to obtain the processed OCT image comprises the following steps:

[0020] S21, performing image denoising processing on the acquired OCT image signal to obtain a first image signal;

[0021] S22, performing contrast enhancement processing on the first image signal to obtain a second image signal;

[0022] S23, performing correction processing on the second image signal to obtain the processed OCT image.

[0023] The correction processing comprises the following steps:

[0024] S231, calculating a gray histogram of each probe element of the OCT image;

[0025] S232, calculating a comprehensive histogram of all probe elements, matching the histogram, and obtaining a relative radiation correction coefficient of each gray value;

[0026] S233, for each pixel point of the second image signal, multiplying the relative radiation correction coefficient corresponding to the gray value of the pixel point by the gray value of the pixel point to obtain a processed OCT image.

[0027] The processed OCT image is analyzed and recognized to obtain real-time position information of the PICC catheter, and the real-time position information of the PICC catheter is compared with a preset optimal position to obtain a position discrimination result, including:

[0028] S41, obtaining three-dimensional image information of the PICC catheter; the three-dimensional image information of the PICC catheter is represented as a three-dimensional matrix S.

[0029] S42, matching and calculating the processed OCT image and the three-dimensional image information of the PICC catheter to obtain a real-time position information sequence of the PICC catheter.

[0030] S43, performing similarity discrimination processing on the real-time position information sequence of the PICC catheter and the preset optimal position to obtain a position discrimination result.

[0031] The matching and calculating the processed OCT image and the three-dimensional image information of the PICC catheter to obtain a real-time position information sequence of the PICC catheter, including:

[0032] S421, performing three-dimensional convolution calculation on the processed OCT image and the three-dimensional image information of the PICC catheter at each time to obtain a convolution value of each three-dimensional position of the processed OCT image.

[0033] S422, determining a three-dimensional position with the largest convolution value of the processed OCT image as the real-time position information of the PICC catheter at the time.

[0034] S423, constructing a real-time position information sequence of the PICC catheter by using the real-time position information of the PICC catheter at all times.

[0035] The similarity discrimination processing on the real-time position information sequence of the PICC catheter and the preset optimal position to obtain a position discrimination result, including:

[0036] S431, performing total similarity value calculation processing on the real-time position information sequence of the PICC catheter and the preset optimal position to obtain a total similarity value.

[0037] S432, determining whether the total similarity value is greater than a set similarity threshold, obtaining a first discrimination result; if the first discrimination result is yes, determining that the position discrimination result is that the preset position is reached; if the first discrimination result is no, determining that the position discrimination result is that the preset position is not reached.

[0038] The expression of the total similarity value calculation process is:

[0039]

[0040] Wherein, R(t) is the position information at time t in the real-time position information sequence, Q is the preset optimal position, P is the total similarity value, E is the total number of times, and || represents the Euclidean distance calculation.

[0041] The third aspect of the present application discloses a PICC catheter positioning and identifying device based on optical coherence tomography, which comprises:

[0042] A memory storing executable program codes;

[0043] A processor coupled with the memory;

[0044] The processor calls the executable program codes stored in the memory to execute the PICC catheter positioning and identifying method based on optical coherence tomography.

[0045] The fourth aspect of the present application discloses a computer storage medium storing computer instructions, which are called by a computer to execute the PICC catheter positioning and identifying method based on optical coherence tomography.

[0046] The fifth aspect of the present application discloses an information data processing terminal for implementing the PICC catheter positioning and identifying method based on optical coherence tomography.

[0047] The present application has the following advantages:

[0048] High-resolution imaging: OCT technology can provide high-resolution intravascular images, clearly showing the relative position of the catheter and the blood vessel wall, and improving the accuracy of catheterization.

[0049] Real-time imaging: The device can acquire and display OCT images in real time, helping medical staff to monitor the catheter insertion process in real time.

[0050] Portability: The device is designed to be light and portable, making it easy to use in different medical scenarios and improving medical efficiency.

[0051] Intelligent alarm: the combination of position recognition unit and alarm unit can timely remind medical staff whether the catheter reaches the optimal position, and reduce operation errors.

[0052] Real-time position detection and high detection accuracy: the application proposes corresponding matching calculation processing and similarity discrimination processing algorithms for target recognition and position detection of OCT images, obtains real-time position information sequence of the PICC catheter by matching calculation processing of the processed OCT image and three-dimensional image information of the PICC catheter, and obtains position discrimination results by similarity discrimination processing of the real-time position information sequence of the PICC catheter and the preset optimal position, thereby improving the detection accuracy.

[0053] The application establishes a corresponding image correction model for the characteristics of OCT images before image recognition and detection, ensures the accuracy of the collected images, and provides image sources for correct detection. BRIEF DESCRIPTION OF DRAWINGS

[0054] Figure 1 It is a schematic diagram of the device of the application;

[0055] Figure 2 It is a flowchart of the method of the application. DETAILED DESCRIPTION

[0056] In order to better understand the content of the application, an embodiment is given here.

[0057] Figure 1 It is a schematic diagram of the device of the application; Figure 2 It is a flowchart of the method of the application.

[0058] In the first aspect of the embodiment of the application, a PICC catheter positioning and recognition device based on optical coherence tomography (OCT) is disclosed, which comprises an OCT probe, a signal acquisition and processing unit, an image display unit and a position recognition unit.

[0059] The OCT probe is used to obtain OCT images of blood vessels and PICC catheters in the PICC catheterization process; the probe has high resolution and can scan the internal structure of blood vessels in real time, clearly showing the relative position of the catheter and the blood vessel wall.

[0060] The signal acquisition and processing unit is connected with the OCT probe and is used to process the acquired OCT image signals to obtain processed OCT images.

[0061] The image display unit is connected with the signal acquisition and processing unit and is used to display the processed OCT images in real time for medical staff to observe.

[0062] The position recognition unit is connected with the image display unit and the signal acquisition and processing unit, and is configured to analyze and recognize the processed OCT image to obtain real-time position information of the PICC catheter, compare the real-time position information with a preset optimal position, and obtain a position discrimination result.

[0063] The OCT probe is implemented by using a cardiovascular OCT imaging device; specifically, a cardiovascular OCT system of a micro-light medical device and a disposable intravascular imaging catheter, or an intravascular OCT imaging system and an imaging catheter of a constant medical device, or an OCT device and an imaging catheter of a Wofu medical device can be used to implement the OCT probe.

[0064] The signal acquisition and processing unit is configured to process the acquired OCT image signal to obtain a processed OCT image, and the processed OCT image comprises:

[0065] The signal acquisition and processing unit performs image denoising processing on the acquired OCT image signal to obtain a first image signal.

[0066] The first image signal is subjected to contrast enhancement processing to obtain a second image signal.

[0067] The second image signal is subjected to correction processing to obtain the processed OCT image.

[0068] The correction processing comprises:

[0069] A gray level histogram of each probe element of the OCT image is calculated.

[0070] A comprehensive histogram of all probe elements is calculated, the histogram is matched, and a relative radiation correction coefficient of each gray level value is obtained.

[0071] For each pixel point of the second image signal, the relative radiation correction coefficient corresponding to the gray level value of the pixel point is multiplied by the gray level value of the pixel point to obtain the processed OCT image.

[0072] The position recognition unit analyzes and recognizes the processed OCT image to obtain real-time position information of the PICC catheter, compares the real-time position information of the PICC catheter with a preset optimal position, and obtains a position discrimination result, which comprises:

[0073] Three-dimensional image information of the PICC catheter is obtained; the three-dimensional image information of the PICC catheter is represented as a three-dimensional matrix S.

[0074] The processed OCT image and the three-dimensional image information of the PICC catheter are subjected to matching calculation processing to obtain a real-time position information sequence of the PICC catheter.

[0075] The real-time position information sequence of the PICC catheter is compared with the preset optimal position to obtain a position discrimination result.

[0076] The matching calculation of the processed OCT image and the three-dimensional image information of the PICC catheter is performed to obtain the real-time position information sequence of the PICC catheter, including:

[0077] The processed OCT image at each time is three-dimensionally convoluted with the three-dimensional image information of the PICC catheter to obtain a convolution value of each three-dimensional position of the processed OCT image.

[0078] The three-dimensional position of the processed OCT image with the largest convolution value is determined as the real-time position information of the PICC catheter at the time.

[0079] The real-time position information sequence of the PICC catheter is constructed by using the real-time position information of the PICC catheter at all times.

[0080] The real-time position information sequence of the PICC catheter is compared with the preset optimal position to obtain a position discrimination result, including:

[0081] The real-time position information sequence of the PICC catheter is compared with the preset optimal position to obtain a position discrimination result, including:

[0082] It is judged whether the total similarity value is greater than a set similarity threshold to obtain a first discrimination result; if the first discrimination result is yes, it is determined that the position discrimination result is that the preset position is reached; if the first discrimination result is no, it is determined that the position discrimination result is that the preset position is not reached.

[0083] The expression of the total similarity value calculation processing is:

[0084]

[0085] wherein R(t) is the position information at time t in the real-time position information sequence, Q is the preset optimal position, P is the total similarity value, E is the total number of times, and || represents the Euclidean distance calculation.

[0086] When the position discrimination result is that the preset position is reached, the position recognition unit is connected to the alarm unit to send an alarm signal, and the alarm unit sends an alarm information.

[0087] The correction processing includes:

[0088] For the jth column of the OCT probe, the gray histogram P of the image obtained by the single probe element is calculated j and the cumulative probability density function S j, 1≤j≤N;

[0089] For the gray histogram P j , the corresponding cumulative probability density function S j (k) is:

[0090] P j (k) = m j (k) / M j ,

[0091] In the formula, m j (k) is the number of pixels with the gray value k in the image obtained by the jth detector, M j is the total number of pixels in the image obtained by the jth detector, and then for the gray histogram P j , the corresponding cumulative probability density function S j (k) is:

[0092]

[0093] wherein, l is the gray value of the pixel; the gray histograms of each detector are combined to obtain a comprehensive histogram P of all detectors; for the comprehensive histogram P, the calculation formula of the corresponding comprehensive histogram P(k) when the gray value is k is:

[0094]

[0095] For the comprehensive histogram P of all detectors, the corresponding cumulative probability density function V(k) when the gray value of the pixel is k is:

[0096]

[0097] In the histogram matching process, the comprehensive histogram of all detectors is the expected histogram, and a lookup table is established according to the expected histogram, so that the probability density function of the comprehensive histogram of each detector after the matching processing is the same as the probability density function of the expected histogram;

[0098] According to the corresponding cumulative probability density function S j of the gray histogram P j of the jth detector and the corresponding cumulative probability density function V of the expected histogram, the relative radiation correction coefficient of the jth detector is calculated, and the relative radiation correction coefficient of the jth detector is recorded in the gray lookup table T j , 1≤j≤N; T j represents the gray lookup table corresponding to the jth detector.

[0099] For the cumulative probability density function S j of the gray histogram P j(k), starting from the gray value 0, search for the gray value I in the cumulative probability density function V corresponding to the expected histogram, which satisfies the following condition:

[0100] V(I)≤S j (k)≤V(I+1),

[0101] After finding the gray value I, if |V(I)-S j (k)|-|V(I+1)-S j (k)|≤0, the gray lookup table T j The relative radiation correction coefficient T j (k) recorded at the pixel gray value k for the jth probe element is:

[0102] T j (k)=I,

[0103] If |V(I)-S j (k)|-|V(I+1)-S j (k)|>0, the gray lookup table T j The relative radiation correction coefficient T j (k) recorded at the pixel gray value k for the jth probe element is:

[0104] T j (k)=I+1,

[0105] For each probe element, the above operation process is performed to obtain the relative radiation correction coefficient for each gray value.

[0106] For each pixel point of the second image signal, the relative radiation correction coefficient corresponding to the gray value of the pixel point is multiplied by the gray value of the pixel point to obtain a processed OCT image.

[0107] The expression of the three-dimensional convolution calculation is:

[0108] Con(i,j,k)=conv(R(i,j,k),S),

[0109] where R(i,j,k) represents the pixel point of the jth row and the kth column of the i th layer of the processed OCT image, Con(i,j,k) represents the convolution value of the pixel point of the jth row and the kth column of the i th layer of the OCT image, and the three-dimensional position of the pixel point of the jth row and the kth column of the i th layer of the OCT image is (i,j,k).

[0110] The three-dimensional convolution calculation can be realized by using a three-dimensional convolution module in the neural network, and specifically can be realized by using torch.nn.Conv3d in the PyTorch framework, and torch.nn.Conv3d is a module specially used for realizing three-dimensional convolution in the PyTorch. The three dimensions of depth, height and width of the convolution kernel are set to apply the convolution operation on the three-dimensional volume of the input data. For example, torch.nn.Conv3d(in_channels, out_channels, kernel_size, stride=1, padding=0, dilation=1, groups=1, bias=True, padding_mode='zeros'), wherein in_channels is the number of input channels, out_channels is the number of output channels, kernel_size is the size of the convolution kernel, can be an integer or a three-tuple to specify the size of the depth, height and width directions respectively, stride is the step length of the convolution kernel, padding is the size of the padding around the input data, dilation is the spacing between the elements of the convolution kernel, groups is the number of groups of the grouped convolution, bias is whether to add a bias term, and padding_mode is the padding mode.

[0111] The processed OCT image of a moment is a processed OCT image obtained at a collection moment.

[0112] The device further comprises an alarm unit connected with the position recognition unit, which gives an audible and visual alarm prompt when the PICC catheter reaches the preset optimal position.

[0113] During the PICC catheterization operation, the OCT probe is placed on the patient's body surface or inserted into the blood vessel through an adapter to obtain real-time OCT images of the blood vessel and the PICC catheter. After the image processing unit processes the images, the image display unit displays the processed images in real time. The position recognition unit analyzes the images to identify the catheter position, and the alarm unit gives an alarm prompt when the catheter reaches the optimal position, helping medical staff accurately complete the catheterization operation.

[0114] In this embodiment, the OCT probe adopts a handheld design, which is convenient for medical staff to operate. The signal acquisition and processing unit is integrated in a portable device and connected with the OCT probe through wireless or wired mode. The image display unit is a high-resolution touch screen that displays OCT images in real time and provides interactive functions such as image zooming and rotating. The position recognition unit uses advanced image recognition algorithms to quickly and accurately identify the catheter position. The alarm unit includes sound alarm and light flashing to ensure that medical staff can promptly notice the alarm prompt.

[0115] In the embodiment, the OCT probe is designed as a disposable sterile probe suitable for PICC catheterization operation of different patients. The signal acquisition and processing unit and the image display unit are integrated in a portable medical device, and the device shell is designed with waterproof and dustproof to be suitable for use in various medical environments. The position recognition unit combines artificial intelligence technology to improve the accuracy of catheter position recognition through a deep learning algorithm. The alarm unit increases the vibration alarm function in addition to the sound and light alarm to adapt to the needs of different medical staff.

[0116] In the second aspect of the embodiment, a PICC catheter positioning and recognition method based on optical coherence tomography (OCT) is disclosed, which is realized by using the PICC catheter positioning and recognition device based on optical coherence tomography (OCT), and includes the following steps:

[0117] An OCT probe is used to obtain OCT images of blood vessels and PICC catheters in the PICC catheterization process.

[0118] A signal acquisition and processing unit is used to process the collected OCT image signals to obtain processed OCT images.

[0119] An image display unit is used to display the processed OCT images in real time for medical staff to observe.

[0120] A position recognition unit is used to analyze and identify the processed OCT images to obtain real-time position information of the PICC catheter, and the real-time position information of the PICC catheter is compared with a preset optimal position to obtain a position discrimination result.

[0121] The processing of the collected OCT image signals to obtain the processed OCT images includes the following steps:

[0122] The collected OCT image signals are subjected to image denoising processing to obtain first image signals.

[0123] The first image signals are subjected to contrast enhancement processing to obtain second image signals.

[0124] The second image signals are subjected to correction processing to obtain the processed OCT images.

[0125] The correction processing includes the following steps:

[0126] The gray level histogram of each probe element of the OCT image is calculated.

[0127] The comprehensive histogram of all probe elements is calculated, the histogram is matched, and the relative radiation correction coefficient of each gray value is obtained.

[0128] For each pixel point of the second image signal, a relative radiation correction coefficient corresponding to a gray value of the pixel point is multiplied by the gray value to obtain a processed OCT image;

[0129] The processed OCT image is analyzed and recognized to obtain real-time position information of the PICC catheter, and the real-time position information of the PICC catheter is compared with a preset optimal position to obtain a position discrimination result, including:

[0130] Three-dimensional image information of the PICC catheter is obtained; the three-dimensional image information of the PICC catheter is represented as a three-dimensional matrix S;

[0131] The processed OCT image and the three-dimensional image information of the PICC catheter are matched and calculated to obtain a real-time position information sequence of the PICC catheter;

[0132] The real-time position information sequence of the PICC catheter and the preset optimal position are similarity discriminated to obtain a position discrimination result;

[0133] The three-dimensional image information of the PICC catheter is obtained by using an OCT probe to capture the three-dimensional image information of the PICC catheter;

[0134] The processed OCT image and the three-dimensional image information of the PICC catheter are matched and calculated to obtain a real-time position information sequence of the PICC catheter, including:

[0135] For each time, the processed OCT image and the three-dimensional image information of the PICC catheter are three-dimensionally convoluted to obtain a convolution value of each three-dimensional position of the processed OCT image;

[0136] A three-dimensional position of the processed OCT image with the largest convolution value is determined as real-time position information of the PICC catheter at the time;

[0137] The real-time position information sequence of the PICC catheter is constructed by using real-time position information of the PICC catheter at all times.

[0138] The real-time position information sequence of the PICC catheter and the preset optimal position are similarity discriminated to obtain a position discrimination result, including:

[0139] The real-time position information sequence of the PICC catheter and the preset optimal position are similarity discriminated to obtain a position discrimination result, including:

[0140] determining whether the total similarity value is greater than a set similarity threshold to obtain a first discrimination result; if the first discrimination result is yes, determining that the position discrimination result is that the preset position is reached; if the first discrimination result is no, determining that the position discrimination result is that the preset position is not reached.

[0141] The expression of the total similarity value calculation processing is:

[0142]

[0143] wherein R(t) is the position information at time t in the real-time position information sequence, Q is the preset optimal position, P is the total similarity value, E is the total number of times, and || represents the Euclidean distance calculation.

[0144] When the position discrimination result is that the preset position is reached, the position recognition unit is connected to the alarm unit to send an alarm signal, and the alarm unit sends an alarm information.

[0145] The correction processing comprises:

[0146] For the jth column of the OCT probe, the gray histogram P of the image obtained by the single probe element is calculated j and the cumulative probability density function S j , 1≤j≤N.

[0147] For the gray histogram P j , when the gray value is k, the calculation formula of the corresponding gray histogram P j (k) is:

[0148] P j (k) = m j (k) / M j ,

[0149] In the formula, m j (k) is the number of pixels with a gray value equal to k in the image obtained by the jth probe element, M j is the total number of pixels in the image obtained by the jth probe element, and then for the gray histogram P j , when the pixel gray value is k, the corresponding cumulative probability density function S j (k) is:

[0150]

[0151] wherein l is the gray value of the pixel; the gray histograms of each probe element are merged to obtain a comprehensive histogram P of all the probe elements; for the comprehensive histogram P, when the gray value is k, the calculation formula of the corresponding comprehensive histogram P(k) is:

[0152]

[0153] For the comprehensive histogram P of all the detectors, the corresponding cumulative probability density function V(k) at the pixel gray value k is:

[0154]

[0155] In the histogram matching process, the comprehensive histogram of all the detectors is the expected histogram, and a lookup table is established according to the expected histogram, so that the probability density function of the comprehensive histogram of each detector after the matching process is the same as the probability density function of the expected histogram;

[0156] According to the gray histogram Pj of the jth detector j The corresponding cumulative probability density function Sj j And the cumulative probability density function V of the expected histogram, the relative radiation correction coefficient of the jth detector is calculated, and the relative radiation correction coefficient of the jth detector is recorded in the gray lookup table T j , 1≤j≤N; T j represents the gray lookup table corresponding to the jth detector;

[0157] For the cumulative probability density function Sj j (k) of the pixel gray value k in the gray histogram Pj j (k), starting from the gray value 0, search for the gray value l in the cumulative probability density function V of the expected histogram that satisfies the following condition:

[0158] V(l)≤S j (k)≤V(l+1),

[0159] After finding the gray value l, when |V(l)-S j (k)|-|V(l+1)-S j (k)|≤0, the relative radiation correction coefficient T j recorded in the gray lookup table T j of the jth detector at the pixel gray value k is:

[0160] T j (k)=l,

[0161] When |V(l)-S j (k)|-|V(l+1)-S j (k)|>0, the relative radiation correction coefficient T j recorded in the gray lookup table T j of the jth detector at the pixel gray value k is:

[0162] T j (k)=l+1,

[0163] The above operation process is performed for each probe element, and the relative radiation correction coefficient of each gray value is obtained.

[0164] For each pixel point of the second image signal, the relative radiation correction coefficient corresponding to the gray value of the pixel point is multiplied by the gray value of the pixel point to obtain a processed OCT image.

[0165] The expression of the three-dimensional convolution calculation is:

[0166] Con(i,j,k)=conv(R(i,j,k),S),

[0167] wherein R(i,j,k) represents a pixel point of the i-th layer, the j-th row and the k-th column of the processed OCT image, Con(i,j,k) represents a convolution value of a pixel point of the i-th layer, the j-th row and the k-th column of the OCT image, and the three-dimensional position of the pixel point of the i-th layer, the j-th row and the k-th column of the OCT image is (i,j,k).

[0168] The three-dimensional convolution calculation can be realized by using a three-dimensional convolution module in a neural network, and specifically can be realized by using torch.nn.Conv3d in a PyTorch framework, which is a module specially used for realizing three-dimensional convolution in PyTorch. It applies a convolution operation on a three-dimensional volume of input data by setting parameters of three dimensions of depth, height and width of a convolution kernel. For example, torch.nn.Conv3d(in_channels, out_channels, kernel_size, stride=1, padding=0, dilation=1, groups=1, bias=True, padding_mode='zeros'), wherein in_channels is the number of input channels, out_channels is the number of output channels, kernel_size is the size of the convolution kernel, which can be an integer or a three-tuple to respectively specify the size of the depth, height and width directions, stride is the step length of the convolution kernel, padding is the size of the padding around the input data, dilation is the spacing between the elements of the convolution kernel, groups is the number of groups of the grouped convolution, bias is whether to add a bias term, and padding_mode is the padding mode.

[0169] The above only describes the embodiments of the present application and is not used to limit the present application. The present application can have various changes and variations for those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the scope of claims of the present application.

Claims

1. An optical coherence tomography-based PICC catheterization positioning and recognition device, characterized in that, The application relates to an OCT (Optical Coherence Tomography) real-time position identification method and device for PICC (Peripherally Inserted Central Catheter) catheterization. The OCT probe, the signal acquisition and processing unit, the image display unit and the position identification unit are comprised; The OCT probe is used for acquiring OCT images of blood vessels and a PICC catheter in a PICC catheterization process; The signal acquisition and processing unit is connected with the OCT probe and is used for processing the acquired OCT image signals to obtain processed OCT images; The image display unit is connected with the signal acquisition and processing unit and is used for displaying the processed OCT images in real time; The position identification unit is connected with the image display unit and the signal acquisition and processing unit and is used for analyzing and identifying the processed OCT images to obtain real-time position information of the PICC catheter, comparing the real-time position information of the PICC catheter with a preset optimal position to obtain a position identification result, which comprises the following steps: S41, three-dimensional image information of the PICC catheter is obtained; the three-dimensional image information of the PICC catheter is expressed as a three-dimensional matrix S; S42, the processed OCT images and the three-dimensional image information of the PICC catheter are matched and calculated to obtain a real-time position information sequence of the PICC catheter; S43, the real-time position information sequence of the PICC catheter and the preset optimal position are subjected to similarity discrimination processing to obtain a position identification result, which comprises the following steps: S431, total similarity value calculation processing is performed on the real-time position information sequence of the PICC catheter and the preset optimal position to obtain a total similarity value; S432, whether the total similarity value is greater than a set similarity threshold value is judged to obtain a first discrimination result; if the first discrimination result is yes, it is determined that the position identification result is that the preset position is reached; if the first discrimination result is no, it is determined that the position identification result is that the preset position is not reached; The expression of the total similarity value calculation processing is as follows: Wherein, R(t) is position information at t moment in the real-time position information sequence, Q is the preset optimal position, P is the total similarity value, E is the total number of moments, and || represents the Euclidean distance calculation.

2. The optical coherence tomography based PICC catheterization positioning and identification device of claim 1, wherein, The OCT probe is realized by using a cardiovascular OCT imaging device.

3. The OCT-based PICC catheter placement identification device of claim 1, wherein, The signal acquisition and processing unit is used for processing the acquired OCT image signals to obtain processed OCT images, which comprises the following steps: The signal acquisition and processing unit performs image denoising processing on the acquired OCT image signals to obtain a first image signal; The first image signal is subjected to contrast enhancement processing to obtain a second image signal; The second image signal is subjected to correction processing to obtain the processed OCT images.

4. The optical coherence tomography based PICC catheterization positioning and identification device of claim 1, wherein, The matching and calculation processing of the processed OCT images and the three-dimensional image information of the PICC catheter to obtain the real-time position information sequence of the PICC catheter comprises the following steps: S421, three-dimensional convolution calculation is performed on the processed OCT images at each moment and the three-dimensional image information of the PICC catheter to obtain convolution values of each three-dimensional position of the processed OCT images; S422, the three-dimensional position with the largest convolution value of the processed OCT images is determined as the real-time position information of the PICC catheter at the moment; S423, using the real-time position information of the PICC catheter at all times, a real-time position information sequence of the PICC catheter is constructed.

Images