Parameter determination method and device, computer device, and storage medium

Abstract

The application relates to a parameter determination method and device, computer equipment and a storage medium. The method comprises the following steps: obtaining a plurality of parameter curves of a region of interest based on a neck transparent layer in a target medical image, and extracting a target region in the region of interest according to a wave crest and a wave trough in each parameter curve, so as to determine the thickness of the neck transparent layer according to the target region. The method can improve the accuracy.

Description

Technical Field

[0001] This application relates to the field of imaging technology, and in particular to a parameter determination method, apparatus, computer device, and storage medium. Background Technology

[0002] The thickness of the nuchal translucency (NT) plays a significant role in the screening of fetal chromosomal abnormalities; therefore, determining the thickness of the NT is an important task.

[0003] Currently, doctors typically manually draw the nuchal translucency region in the target image and determine the NT thickness based on experience. However, the current methods for determining these parameters rely too heavily on the doctor's experience. Summary of the Invention

[0004] Therefore, it is necessary to provide a parameter determination method, apparatus, computer equipment, and storage medium that reduces reliance on doctors to address the aforementioned technical problems.

[0005] Firstly, this application provides a parameter determination method, including:

[0006] Based on the region of interest in the neck transparency layer of the target medical image, multiple parameter curves of the region of interest are obtained;

[0007] Based on the peaks and troughs in the curves of each parameter, extract the target region within the region of interest;

[0008] The thickness of the neck translucency layer is determined based on the target area.

[0009] Secondly, this application also provides a parameter determining device, comprising:

[0010] The first determining module is used to obtain multiple parameter curves of the region of interest based on the region of interest in the neck transparency layer of the target medical image.

[0011] The extraction module is used to extract the target region within the region of interest based on the peaks and troughs in the curves of each parameter.

[0012] The second determining module is used to determine the thickness of the neck transparency layer based on the target area.

[0013] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of any of the above methods.

[0014] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of any of the above methods.

[0015] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of any of the above methods.

[0016] The aforementioned parameter determination method, apparatus, computer equipment, and storage medium can obtain multiple parameter curves for the region of interest (ROI) of the neck transparency layer in a target medical image. Based on the peaks and troughs of each parameter curve, a target region within the ROI is extracted to determine the thickness of the neck transparency layer. This reduces reliance on physicians in the thickness determination process, thereby minimizing the influence of physician experience on the neck transparency layer thickness and preventing significant discrepancies in the thickness determined by different physicians for the same target medical image, thus improving the consistency of the determined neck transparency layer thickness. Furthermore, since the target region is extracted from the ROI of the neck transparency layer based on the peaks and troughs of each parameter curve, the characteristics of the neck transparency layer on the parameter curves are taken into account during the target region determination process. On the one hand, this narrows down the target region, increasing the processing speed based on the target region and thus improving the efficiency of determining the neck transparency layer thickness. On the other hand, it improves the precision of the target region, thereby increasing the accuracy of the neck transparency layer thickness determined based on the target region. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments or related technologies of this application, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a diagram illustrating the application environment of the parameter determination method in the embodiments of this application;

[0019] Figure 2 This is a flowchart illustrating the parameter determination method in an embodiment of this application;

[0020] Figure 3 This is a schematic diagram illustrating a process for determining a region of interest in an embodiment of this application;

[0021] Figure 4 This is a schematic diagram of a parameter curve in an embodiment of this application;

[0022] Figure 5This is a schematic diagram of a target area in an embodiment of this application;

[0023] Figure 6 This is a schematic diagram illustrating a process for determining the thickness of the neck transparency layer in an embodiment of this application;

[0024] Figure 7 This is a schematic diagram of a process for determining a target area in an embodiment of this application;

[0025] Figure 8 This is a schematic diagram illustrating a process for obtaining a target region in an embodiment of this application;

[0026] Figure 9 This is a schematic diagram of a process for determining a discrimination score in an embodiment of this application;

[0027] Figure 10 This is a schematic diagram of another process for determining the discrimination score in an embodiment of this application;

[0028] Figure 11 This is a schematic diagram of another process for determining a target area in an embodiment of this application;

[0029] Figure 12 This is a schematic diagram of a process for determining a seed image in an embodiment of this application;

[0030] Figure 13 This is a schematic diagram of another process for determining a seed image in an embodiment of this application;

[0031] Figure 14 This is a schematic diagram of another process for determining a seed image in an embodiment of this application;

[0032] Figure 15 This is a schematic diagram illustrating a process for obtaining a seed image in an embodiment of this application;

[0033] Figure 16 This is a schematic diagram of another process for determining the contour of the neck transparency layer in an embodiment of this application;

[0034] Figure 17 This is a schematic diagram of the outline of a neck transparent layer in an embodiment of this application;

[0035] Figure 18 This is a schematic diagram of another process for determining the contour of the neck transparency layer in an embodiment of this application;

[0036] Figure 19 This is a schematic diagram illustrating another process for determining the thickness of the neck transparency layer in an embodiment of this application;

[0037] Figure 20 This is a schematic diagram illustrating another process for determining the thickness of the neck transparency layer in an embodiment of this application;

[0038] Figure 21 This is a schematic diagram of a process for determining a parameter curve in an embodiment of this application;

[0039] Figure 22 This is a schematic diagram of a process for determining a target pixel column in an embodiment of this application;

[0040] Figure 23 This is a visual flowchart illustrating an embodiment of this application;

[0041] Figure 24 This is a schematic diagram of a visual interface in an embodiment of this application;

[0042] Figure 25 This is a flowchart illustrating a parameter determination method according to an embodiment of this application.

[0043] Figure 26 This is a schematic diagram illustrating a parameter determination method in an embodiment of this application;

[0044] Figure 27 This is a schematic diagram illustrating another parameter determination method in the embodiments of this application;

[0045] Figure 28 This is a schematic diagram of the architecture of a parameter determination system according to an embodiment of this application;

[0046] Figure 29 This is a structural block diagram of the parameter determination and adjustment device in the embodiments of this application. Detailed Implementation

[0047] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0048] Figure 1 This is an application environment diagram of the parameter determination method in the embodiments of this application. In an exemplary embodiment, a computer device is provided. This computer device may be a terminal, and its internal structure diagram may be as follows. Figure 1As shown, the computer device includes a processor, memory, input / output interfaces, a communication interface, a display unit, and an input device. The processor, memory, and input / output interfaces are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The input / output interfaces are used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it implements a parameter determination method. The display unit is used to form a visually visible image and can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.

[0049] This embodiment illustrates the method by applying it to a terminal. It is understood that this method can also be applied to a server, and to a system that includes both a terminal and a server, and is implemented through interaction between the terminal and the server. The terminal can be, but is not limited to, various personal computers, laptops, smartphones, and tablets. The server can be a standalone server or a server cluster consisting of multiple servers.

[0050] Those skilled in the art will understand that Figure 1 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0051] Figure 2 This is a flowchart illustrating the parameter determination method in an embodiment of this application. In one exemplary embodiment, such as... Figure 2 As shown, a parameter determination method is provided, which is applied to... Figure 1 The following explanation uses computer equipment as an example, including the following steps S201 to S203.

[0052] S201, based on the region of interest in the neck transparency layer of the target medical image, obtains multiple parameter curves of the region of interest.

[0053] In this embodiment, the computer device can acquire the target medical image and determine the region of interest (ROI) in the nuchal translucency within the target medical image. The target medical image can be an ultrasound image of the fetus in a midsagittal section. The target medical image can be an image transmitted to the computer device from another device, or it can be an image already stored in the computer device.

[0054] Optionally, the computer device can use a preset recognition algorithm to determine the region of interest in the neck translucency layer from the target medical image. This recognition algorithm includes, but is not limited to, Single Shot MultiBox Detector (SSD) and YOLO (You Only Look Once) algorithms.

[0055] Computer devices can also display target medical images in a visualization interface and, in response to a first operation on the displayed target medical image, determine the region of interest (ROI). This first operation includes, but is not limited to, selection and outlining. For example, a doctor can manually draw an ROI box to determine the ROI in the cervical translucency layer of the target medical image.

[0056] Figure 3 This is a schematic diagram illustrating a process for determining a region of interest in an embodiment of this application. Figure 3 (a) shows a target medical image. Figure 3 The white box in (a) represents the region of interest in the target medical image. Figure 3 (b) shows a computer device determining the region of interest based on a target medical image.

[0057] Furthermore, the computer device can determine multiple parametric curves for the region of interest. These parametric curves depict parameter variations within the standard region of interest. The parametric curves can be one of the following: a grayscale curve, a gradient curve, or a pixel value curve. For example, the parametric curve can be a grayscale curve.

[0058] Optionally, the parameter curve can be a parameter curve of the region of interest in a preset direction. For example, such as... Figure 3 As shown, because the transparent layer of the neck has the characteristics of being bright on both sides and dark in the middle, computer equipment can obtain multiple parameter curves of the region of interest in the longitudinal direction.

[0059] Further optional, the computer device can Figure 3(b) shows the region of interest divided into N rows and M columns, and M parametric curves are determined based on the parameters corresponding to the region of interest in each column. Both N and M are numbers greater than 1. The horizontal axis of the parametric curve represents the pixel position of the region of interest, and the vertical axis represents the parameter value. Taking a grayscale curve as an example, the vertical axis represents the grayscale value.

[0060] S202, extract the target region from the region of interest based on the peaks and troughs in the curves of each parameter.

[0061] Because the neck lumen has the characteristic of being bright on both sides and dark in the middle, there will be peaks and troughs in the curves of each parameter. Figure 4 This is a schematic diagram of a parameter curve in an embodiment of this application. Figure 4 A schematic diagram showing multiple grayscale curves of the region of interest in the vertical direction is shown. For example... Figure 4 As shown, the curves for each parameter will have peaks and troughs as the parameters change.

[0062] Furthermore, based on the characteristics of the neck transparency layer, the target region within the region of interest can be extracted according to the peaks and troughs in the parameter curves. Taking the parameter curve as the parameter curve of the region of interest in the vertical direction as an example, since the neck transparency layer has the characteristic of being bright on both sides and dark in the middle, peaks and troughs will exist at similar positions in the parameter curve of the region where the neck transparency layer is located, and the difference between the peaks and troughs is relatively large. Therefore, it is possible to select... Figure 4 For example, computer equipment can determine the target area based on the similarity of the positions of peaks and troughs in all parameter curves, and the area where the difference between two adjacent peaks in each parameter curve is greater than a preset difference.

[0063] Figure 5 This is a schematic diagram of a target area in an embodiment of this application. Please compare it with... Figure 3 (b) and Figure 5 Since the target region is extracted based on the peaks and troughs in the curves of each parameter, the target region can filter out the impurity regions in the region of interest and retain the target region where the true neck transparency layer is located.

[0064] S203, determine the thickness of the neck transparency layer based on the target area.

[0065] In this embodiment, after determining the target area, the computer device can determine the thickness of the nuchal translucency layer based on the target area. Optionally, the computer device can perform threshold segmentation on the target area to identify the dark fluid area among two brighter areas, and determine the thickness of the nuchal translucency layer by calculating the distance between the upper and lower boundaries of the dark fluid area; the computer device can also display the target area in a visualization interface, and the doctor can measure the thickness of the nuchal translucency layer based on the target area, but this embodiment is not limited to this.

[0066] The aforementioned parameter determination method can obtain multiple parametric curves for the region of interest (ROI) of the neck translucency in a target medical image. Based on the peaks and troughs of each parametric curve, a target region within the ROI is extracted to determine the thickness of the neck translucency. This method reduces reliance on physician experience, minimizing the influence of physician experience on the determined thickness and preventing significant discrepancies between thicknesses determined by different physicians for the same target medical image, thus improving the consistency of the determined thickness. Furthermore, since the target region is extracted from the ROI based on the peaks and troughs of each parametric curve, the characteristics of the neck translucency on the parametric curves are taken into account. This narrows the target area, increasing processing speed and efficiency in determining the neck translucency thickness. It also improves the precision of the target region, thereby enhancing the accuracy of the determined thickness.

[0067] Figure 6 This is a schematic diagram of a process for determining the thickness of the neck transparency layer according to an embodiment of this application. In an exemplary embodiment, such as... Figure 6 As shown, S203 includes S601 to S603.

[0068] S601, segment the target region to obtain a seed image.

[0069] In this embodiment, the seed image includes the connected region where the neck transparency layer is located, so as to Figure 5 For example, a seed image can include a dark region in two brighter areas.

[0070] The computer device can segment the target region using a thresholding algorithm to obtain a seed image. Optionally, the thresholding algorithm can be at least one of the following: maximum entropy thresholding algorithm, Otsu's thresholding algorithm, adaptive thresholding algorithm, and fixed thresholding algorithm.

[0071] For example, a computer device can retain pixels in the target area whose grayscale value is less than the preset segmentation threshold and remove pixels in the target area whose grayscale value is not less than the preset segmentation threshold, in order to obtain a seed image.

[0072] S602, determine the contour of the neck transparency layer based on the seed image and the active contour model.

[0073] The active contour model can be a convolutional neural network (CNN), a recurrent neural network (RNN), or other deep learning networks, machine learning networks, etc.

[0074] Active contour models can further refine the segmentation of seed images, and can include Snake segmentation models, level set segmentation algorithms, etc.

[0075] Furthermore, the computer device can determine the contour of the neck transparency layer based on the seed image and the active contour model. For example, the computer device can input the seed image into the active contour model, obtain the output of the active contour model, and then filter and smooth the output of the active contour model to obtain the contour of the neck transparency layer.

[0076] S603, determine the thickness of the nuchal translucency layer based on its contour.

[0077] Furthermore, after determining the outline of the nuchal translucency, its thickness can be determined based on that outline. For example, a computer device can display the outline of the nuchal translucency in a visual interface, allowing a doctor to measure its thickness based on that outline; alternatively, the computer device can determine the thickness of the nuchal translucency by calculating the maximum distance between the upper and lower boundaries of its outline, but this embodiment is not limited thereto.

[0078] In the above embodiments, since the target region is first segmented to obtain a seed image, and then the contour of the neck transparency layer is determined based on the seed image and the active contour model, and the thickness of the neck transparency layer is determined based on the contour of the neck transparency layer, the target region can be coarsely segmented to obtain a seed image, and then further finely segmented based on the seed image and the active contour model, thereby improving the accuracy of the obtained neck transparency layer contour and thus improving the accuracy of the neck transparency layer thickness.

[0079] Figure 7 This is a schematic diagram of a process for determining a target region according to an embodiment of this application. In an exemplary embodiment, such as... Figure 7 As shown, S202 includes S701 to S704.

[0080] S701, based on the peaks and troughs in each parameter curve, extract the peak-trough pairs in each parameter curve; the peak-trough pairs include the first peak, the second peak, and the trough located between the first peak and the second peak.

[0081] In this embodiment, two peaks and an intermediate trough are referred to as a peak-trough pair. That is, a peak-trough pair includes a first peak, a second peak, and a trough located between the first peak and the second peak. It can be understood that the first peak and the second peak are two adjacent peaks.

[0082] Furthermore, based on the peaks and troughs in each parameter curve, peak-trough pairs can be extracted from each parameter curve. Optionally, the computer device can perform second-order difference processing on each parameter curve to extract the peaks and troughs in each parameter curve, thereby determining the peak-trough pairs in each parameter curve.

[0083] The number of peak-to-trough pairs may differ in the curves for each parameter. Figure 4 For example, it can be seen that, Figure 4 The parameter curves in the data all include two peak-to-trough pairs.

[0084] S702, determine the discrimination score of peak-valley pairs in each parameter curve.

[0085] Furthermore, the computer device calculates the discrimination scores for the peak-valley pairs in each parameter curve. That is, each peak-valley pair corresponds to a discrimination score. For example, suppose the region of interest corresponds to parameter curves A, B, and C. Parameter curve A includes peak-valley pairs A1, A2, and A3; parameter curve B includes peak-valley pairs B1, B2, and B3; and parameter curve C includes peak-valley pairs C1 and C2. Then, peak-valley pair A1 corresponds to a discrimination score of 1, peak-valley pair A2 corresponds to a discrimination score of 2, and so on, with peak-valley pair C3 corresponding to a discrimination score of 8.

[0086] The discrimination score is used to filter peak-valley pairs in each parameter curve to retain those that conform to the characteristics of the nuchal lumen from all peak-valley pairs. For example, a computer device can determine a preset peak-valley pair in the parameter curves of the nuchal lumen under normal conditions and compare the difference between the preset peak-valley pair and the peak-valley pairs in each parameter curve to determine the discrimination score of the peak-valley pairs in each parameter curve based on the difference between the preset peak-valley pair and the peak-valley pairs in each parameter curve. In some embodiments, the computer device can input each parameter curve into a scoring model to obtain the discrimination score of the peak-valley pairs in each parameter curve from the scoring model.

[0087] S703 determines the target peak-valley pairs corresponding to each parameter curve based on the discrimination score.

[0088] Optionally, the discrimination score can be determined based on the peak-to-valley ratio of the peak-to-valley pair. The peak-to-valley ratio can be the ratio between the gray value of the valley and the first sum value, which includes the sum of the gray values ​​of the first peak and the second peak. Thus, the smaller the gray value of the valley, the smaller the discrimination score, which better matches the characteristic of the neck lucid layer being bright on both sides and dark in the middle. Therefore, the probability that the region containing the peak-to-valley pair is the region containing the neck lucid layer is higher. Consequently, for each parameter curve, the computer device can select the peak-to-valley pair with the smallest discrimination score as the target peak-to-valley pair.

[0089] In some embodiments, the discrimination score can also be determined based on the distance between peak-valley pairs on different parameter curves, but this embodiment is not limited thereto.

[0090] Continuing with the example above, suppose that for parametric curve A, the discrimination score of peak-valley pair A1 is 1 > the discrimination score of peak-valley pair A3 is 3 > the discrimination score of peak-valley pair A2 is 2. Then, peak-valley pair A2 is the target peak-valley pair for parametric curve A. Similarly, each parametric curve will have a corresponding target peak-valley pair.

[0091] S704. Determine the target area based on the position of the first peak and the position of the second peak corresponding to the target peak and trough of each parameter curve.

[0092] Furthermore, for each parameter curve, the positions of the first peak and the second peak aligned with the target peak-trough can be determined. The position of the first peak indicates its location within the region of interest, and the position of the second peak indicates its location within the region of interest. In other words, the computer device can locate the positions of the first and second peaks aligned with the target peak-trough within the region of interest.

[0093] Figure 8 This is a schematic diagram illustrating a process for obtaining a target region in an embodiment of this application, such as... Figure 8 As shown, taking 5 parameter curves as an example, in Figure 3 Based on (b), the computer equipment determined the target peak-valley pairs located on the five parametric curves, and determined the first and second peaks of the five target peak-valley pairs. Figure 8 The 10 positions are shown as white dots. The white dots in the top row represent the positions of the first peak and the second peak of the five target wave peaks and troughs, respectively.

[0094] Furthermore, based on the positions of the first and second peaks corresponding to the target peaks and troughs of each parameter curve, the computer device can determine the target area. For example, the computer device can divide an initial range based on the positions of the first and second peaks, and then fine-tune this initial range to obtain the target area. Fine-tuning includes, but is not limited to, moving the target area by a first preset distance, which is determined according to requirements.

[0095] In the above embodiments, since the peak-valley pair includes a first peak, a second peak, and a trough located between the first and second peaks, the peak-valley pair in each parameter curve can be extracted based on the peaks and troughs in each parameter curve. Furthermore, since the discrimination score of the peak-valley pair in each parameter curve can be determined, and the target peak-valley pair corresponding to each parameter curve can be determined based on the discrimination score, and then the target region can be determined based on the positions of the first and second peaks in the target peak-valley pair corresponding to each parameter curve, the characteristics of the neck transparency layer on the parameter curve are taken into account during the determination of the target region, thus improving the accuracy of the target region.

[0096] Figure 9 This is a schematic diagram of a process for determining a discrimination score in an embodiment of this application. In an exemplary embodiment, such as... Figure 9 As shown, S702 includes S901 to S904.

[0097] S901 uses the parameter curve with the largest number of peak-to-trough pairs as the reference parameter curve.

[0098] In this embodiment, the computer device uses the parameter curve with the largest number of peak-valley pairs as the reference parameter curve, based on the number of peak-valley pairs. If there are multiple parameter curves with the largest number of peak-valley pairs, the computer device can use any one of them as the reference parameter curve.

[0099] Continuing with the example above, if the number of peak-valley pairs in parameter curve A is greater than the number of peak-valley pairs in parameter curve B, then the computer device can use parameter curve A as a reference parameter curve.

[0100] S902, for each reference peak-valley pair on the reference parameter curve, determine the distance between the reference peak-valley pair and the peak-valley pairs on other parameter curves; the other parameter curves are parameter curves other than the reference parameter curve.

[0101] In this embodiment, for each reference peak-valley pair on the reference parameter curve, the computer device determines the distance between the reference peak-valley pair and peak-valley pairs on other parameter curves. These other parameter curves are those other than the reference parameter curve. That is, for each reference peak-valley pair, the computer device iterates through all peak-valley pairs in the other parameter curves to determine the distance between the reference peak-valley pair and each peak-valley pair in the other parameter curves.

[0102] Continuing the example above, the reference parameter curve is parameter curve A, and the reference peak-valley pairs include peak-valley pairs A1, A2, and A3. For peak-valley pair A1 of parameter curve A, the computer device will determine the distances between peak-valley pair A1 and peak-valley pairs B1, B2, B3, C1, and C2, respectively. Similarly, for peak-valley pair A2 of parameter curve A, the computer device will also determine the distances between peak-valley pair A2 and peak-valley pairs B1, B2, B3, C1, and C2, respectively. The same applies to peak-valley pair A3, which will not be elaborated further here. It should be noted that the letters mentioned above are used to distinguish between each peak-valley pair and each parameter curve, and do not limit the number of peak-valley pairs and parameter curves.

[0103] S903, determine the candidate groups corresponding to each reference peak-valley pair based on the distance; the candidate groups include peak-valley pairs located in each parameter curve.

[0104] In this embodiment, the candidate group includes peak-valley pairs located in each parameter curve; that is, matching peak-valley pairs in each parameter curve are considered as a group. Optionally, for each reference peak-valley pair, the computer device selects the peak-valley pair with the smallest distance from other parameter curves as a candidate group.

[0105] Continuing the example above, for peak-valley pair A1, based on the distances between peak-valley pair A1 and peak-valley pairs B1, B2, and B3 on parameter curve B, the distance between peak-valley pair A1 and peak-valley pair B1 is the smallest, therefore peak-valley pair A1 matches peak-valley pair B1 on parameter curve B. Based on the distances between peak-valley pair A1 and peak-valley pairs C1 and C2 on parameter curve C, the distance between peak-valley pair A1 and peak-valley pair C1 is the smallest, therefore peak-valley pair A1 matches peak-valley pair C1 on parameter curve C. Furthermore, the computer device considers peak-valley pairs A1, B1, and C1 as candidate groups corresponding to peak-valley pair A1.

[0106] Similarly, the computer equipment can also determine the candidate group corresponding to peak-valley pair A2 and the candidate group corresponding to peak-valley pair A3. In other words, the computer equipment can determine a peak-valley pair from each parameter curve to form a candidate group based on the distance between the reference peak-valley pair and peak-valley pairs on other parameter curves.

[0107] It is understandable that, since the number of peak-valley pairs may differ among different candidate parameters, there may be duplicate peak-valley pairs among the candidate groups. For example, the candidate group corresponding to peak-valley pair A1 includes peak-valley pair A1, peak-valley pair B1, and peak-valley pair C1; the candidate group corresponding to peak-valley pair A2 may include peak-valley pair A2, peak-valley pair B2, and peak-valley pair C1; and the candidate group corresponding to peak-valley pair A3 may include peak-valley pair A3, peak-valley pair B3, and peak-valley pair C2. Therefore, both the candidate groups corresponding to peak-valley pair A1 and peak-valley pair A2 can include peak-valley pair C1.

[0108] S904, determine the discrimination score corresponding to each candidate group.

[0109] Furthermore, the computer device can determine the discrimination score corresponding to each candidate group, where the discrimination scores for the peak-valley pairs within each candidate group are the same. For example, the computer device can calculate the degree of matching in the shape, position, and size of the peak-valley pairs in each candidate group, and determine the discrimination score corresponding to each candidate group based on the degree of matching.

[0110] In the above embodiments, the parameter curve with the largest number of peak-valley pairs is used as the reference parameter curve. For each reference peak-valley pair on the reference parameter curve, the distance between the reference peak-valley pair and peak-valley pairs on other parameter curves is determined. Since the other parameter curves are parameter curves other than the reference parameter curve, further, candidate groups corresponding to each reference peak-valley pair are determined based on the distance. Moreover, the candidate groups include peak-valley pairs located in each parameter curve. Therefore, after determining the discrimination score corresponding to each candidate group, the discrimination score of the peak-valley pairs in each parameter curve can be obtained. Furthermore, the matching degree of the peak-valley pairs in each candidate group can be considered in the process of determining the discrimination score.

[0111] Figure 10 This is a schematic diagram of another process for determining the discrimination score in an embodiment of this application. In an exemplary embodiment, such as... Figure 10 As shown, S904 includes S1001 to S1003.

[0112] S1001, for each candidate group, determine the first ratio between the summation result and the largest peak-to-trough ratio; the summation result is the sum of the peak-to-trough ratios of each peak-to-trough pair in the candidate group.

[0113] In this embodiment, the peak-to-valley ratio is determined based on the parameter values ​​of the first peak, the second peak, and the valley in the peak-to-valley pair of the parameter curve. Taking the parameter curve as a grayscale curve as an example, for any peak-to-valley pair, the computer device can determine the peak-to-valley ratio R of the peak-to-valley pair according to the following formula (1). In formula (1), V represents the grayscale value of the valley in the peak-to-valley pair, P1 represents the grayscale value of the first peak in the peak-to-valley pair, and P2 represents the grayscale value of the second peak in the peak-to-valley pair.

[0114] R = V / (P1 + P2) (1)

[0115] Suppose the total number of parametric curves is M, and the number of candidate groups is N. Let the index of the parametric curve be j, and the index of the candidate group be i, where i and j are integers greater than 1. Then j = 1, 2, 3, ..., M, and i = 1, 2, 3, ..., N. It can be understood that the number of candidate groups is equal to the number of all reference peak-trough pairs on the reference parametric curve.

[0116] For example, continuing the above example, suppose the candidate group corresponding to A1 is candidate group 1, the candidate group corresponding to A2 is candidate group 2, and the candidate group corresponding to the peak-valley pair A3 is candidate group 3. Then when i=1 and j=1, it means that the peak-valley pair in the parameter curve A in candidate group 1 is the peak-valley pair A1.

[0117] Furthermore, the maximum peak-to-trough ratio can also be expressed as:

[0118] Among them, R 1j R represents the peak-to-trough ratio of the peak-to-trough pair on the j-th parameter curve of candidate group 1. 2j R represents the peak-to-trough ratio of the peak-to-trough pair on the j-th parameter curve of candidate group 2. Nj This represents the peak-to-valley ratio of the peak-to-valley pair of candidate group N on the j-th parameter curve.

[0119] Furthermore, for each candidate group, a summation result can be determined; that is, for each candidate group, the sum of the peak-to-trough ratios of each peak-to-trough pair in that candidate group is determined, denoted as . Continuing with the example above, for candidate group 1, the summation result of candidate group 1 is the peak-to-trough ratio R of A1. 11 The peak-to-trough ratio R of B1 12 The ratio of peaks and troughs to the peaks and troughs of C1, R 13 The sum of the results. It is understandable that the smaller the sum of the results of candidate group 1, the smaller the gray value of each peak and trough in candidate group 1 relative to the middle trough, which is more consistent with the characteristic of the transparent layer of the neck being bright on both sides and dark in the middle.

[0120] Furthermore, for each candidate group, a first ratio between the summation result and the maximum peak-to-trough ratio can be determined, that is... The summation result is divided by the largest peak-to-trough ratio to normalize the first ratio, so as to avoid the summation result having too much influence on the discrimination score.

[0121] S1002, for each candidate group, determine the second ratio between the distance standard deviation and the maximum distance standard deviation; the distance standard deviation is determined based on the Euclidean distance between each peak-trough pair in the candidate group.

[0122] In this embodiment, the Euclidean distance between a crest-trough pair is the Euclidean distance between the coordinates of the first crest and the coordinates of the second crest in the pair. The Euclidean distance is represented by D. ij For example, the Euclidean distance between the peaks and troughs of candidate group 1 and A1 in the parametric curve A is R. 11 .

[0123] Furthermore, the distance standard deviation is determined based on the Euclidean distance between each peak-trough pair in the candidate group. For example, the computer device can determine the distance standard deviation σ of each candidate group according to the following equation (2). i .in, This represents the average Euclidean distance between peak-valley pairs in each candidate group. It can be understood that the smaller the standard deviation of the distance in a candidate group, the smaller the distance between peak-valley pairs within that group. In other words, the peak-valley pairs are closer together, making this candidate group more consistent with the characteristic of uniform neck transparency thickness compared to other candidate groups.

[0124]

[0125] The maximum standard deviation of the distance, which is also the maximum Euclidean distance between peak and trough pairs in each candidate group, is denoted as max{σ1, σ2, ..., σ... i}

[0126] Furthermore, for each candidate group, a second ratio between the standard deviation of the distance and the largest standard deviation of the distance can be determined, that is... The purpose of dividing the distance standard deviation by the largest distance standard deviation is to normalize the second ratio and avoid the distance standard deviation having too much influence on the discrimination score.

[0127] S1003, determine the discrimination score based on the first ratio and the second ratio.

[0128] In this embodiment, after determining the first ratio and the second ratio, the discrimination score S can be determined based on the first ratio and the second ratio. i Optionally, the computer device may use the sum of the first ratio and the second ratio as the discrimination score S corresponding to the candidate group. iAs shown in equation (3). Thus, after determining the discrimination scores corresponding to the candidate groups, the candidate groups that are close in position on each parameter curve and have a large difference in distance between the first and second peaks can be identified as the target peak-valley pairs. Discrimination score S i The smaller the value, the more likely the peak-trough pairs in the corresponding candidate group are located in the neck transparency layer.

[0129]

[0130] In the above embodiments, since the summation result is the sum of the peak-to-valley ratios of each peak-to-valley pair in the candidate group, and the distance standard deviation is determined based on the Euclidean distance of each peak-to-valley pair in the candidate group, after determining the first ratio between the summation result and the largest peak-to-valley ratio for each candidate group, and after determining the second ratio between the distance standard deviation and the largest distance standard deviation for each candidate group, the discrimination score can be determined based on the first ratio and the second ratio.

[0131] Figure 11 This is a schematic diagram of another process for determining a target region in an embodiment of this application. In an exemplary embodiment, such as... Figure 11 As shown, S704 includes S1101 to S1103.

[0132] S1101, the first curve is obtained by fitting the position of the first peak in the target peak-valley pair corresponding to each parameter curve.

[0133] In this embodiment, we continue to use Figure 8 For example, computer equipment can... Figure 8 The first curve is obtained by fitting the dots in the top row.

[0134] S1102, the second curve is obtained by fitting the position of the second peak in the target peak-valley pair corresponding to each parameter curve.

[0135] Similarly, computer equipment can... Figure 8 The second curve is obtained by fitting the dots in the lower row.

[0136] S1103, Determine the target region based on the intersection points of the first curve and the second curve with the boundary of the region of interest.

[0137] Furthermore, since the first curve and the second curve can intersect with the boundary of the region of interest, the computer device can determine the target region based on the intersection points of the first curve and the second curve with the boundary of the region of interest, respectively.

[0138] For example, the computer device can move the intersection points of the first curve and the second curve with the boundary of the region of interest upward or downward by a second preset distance, and crop the region of interest based on the moved intersection points to obtain the target region. The obtained target region can be as follows: Figure 5 As shown. In some embodiments, the computer device may also filter the cropped region after cropping the region of interest to obtain the target region. The second preset distance can be determined based on the thickness between the first and second curves and the height of the target medical image.

[0139] In the above embodiments, since the first curve is obtained by fitting the position of the first peak in the target peak-valley pair corresponding to each parameter curve, and the second curve is obtained by fitting the position of the second peak in the target peak-valley pair corresponding to each parameter curve, the target region can be determined relatively accurately based on the intersection points of the first curve and the second curve with the boundary of the region of interest.

[0140] Figure 12 This is a schematic diagram of a process for determining a seed image according to an embodiment of this application. In an exemplary embodiment, such as... Figure 12 As shown, S601 includes S1201 to S1203.

[0141] S1201, The target region is segmented according to a preset threshold segmentation algorithm to obtain the first image.

[0142] The preset threshold segmentation algorithm may include, but is not limited to, at least one of the following: high entropy threshold segmentation algorithm, Otsu's threshold segmentation algorithm, adaptive threshold segmentation algorithm, and fixed threshold segmentation algorithm.

[0143] For example, the preset threshold segmentation algorithm can be the maximum entropy segmentation method. Furthermore, the computer device determines a binarization threshold using the grayscale probability information corresponding to the target region, and uses this binarization threshold to segment the target region, using the binarized image smaller than the binarization threshold as the first image.

[0144] In some embodiments, the computer device may also filter the target region and then segment the filtered target region to obtain a first image. It is understood that the first image includes the connected region containing the neck transparency layer.

[0145] S1202, Perform a first process on the first image to obtain a second image.

[0146] The first processing may include filtering, smoothing, or other methods. For example, a computer device may smooth the first image to obtain a second image.

[0147] S1203, if the second image meets the preset quality conditions, the second image is used as the seed image.

[0148] Furthermore, the computer device can filter the second image using preset quality conditions. If the second image meets the preset quality conditions, it can then be used as the seed image.

[0149] The preset quality conditions can be determined according to requirements. For example, if the aspect ratio of the connected components in the second image is within the preset aspect ratio range, then the preset quality conditions are met; otherwise, they are not met.

[0150] In the above embodiments, the target region is first segmented according to a preset threshold segmentation algorithm to obtain a first image, and then the first image is processed to obtain a second image. Since the second image is used as a seed image only if it meets preset quality conditions, the accuracy of the seed image is improved.

[0151] In one exemplary embodiment, optionally, the first process includes at least one of morphological erosion processing, area filtering, and centroid location filtering. Specifically, area filtering includes retaining connected components in the first image with an area greater than a preset area. Centroid location filtering includes retaining connected components in the first image whose centroids are within a preset centroid range.

[0152] In this embodiment, the first image obtained by segmenting the target region according to the preset threshold segmentation algorithm may contain multiple connected regions of different sizes and locations. Most of these connected regions are impurity regions. In order to preserve the connected region where the neck transparency layer is located, this embodiment will perform at least one of the following on the first image: morphological erosion processing, area filtering, and centroid position filtering.

[0153] Morphological erosion can remove connected components that are stuck together in the first image. Based on the characteristics of the neck transparency layer, area filtering is used to retain connected components in the first image with an area greater than a preset area. Centroid position filtering is used to retain connected components in the first image whose centroids are within a preset centroid range.

[0154] For example, after performing morphological erosion on the first image, the computer device calculates the area of ​​each connected component in the first image and retains the connected component with the largest area. Then, the computer device calculates the centroids of the remaining connected components in the first image and retains the connected components whose centroids fall within a preset centroid range.

[0155] In this embodiment, since the first processing includes at least one of morphological erosion processing, area filtering, and centroid position filtering, and area filtering includes retaining connected components in the first image whose area is greater than a preset area; and centroid position filtering includes retaining connected components in the first image whose centroid is within a preset centroid range, a more accurate second image can be obtained after performing the first processing on the first image.

[0156] Figure 13 This is a schematic diagram of another process for determining a seed image in an embodiment of this application. In an exemplary embodiment, such as... Figure 13 As shown, the above parameter determination method also includes S1301 to S1303.

[0157] S1301, if the second image does not meet the preset quality conditions, determine the seed point according to the position of the first peak and the position of the second peak in the target peak-valley alignment corresponding to each parameter curve.

[0158] In this embodiment, if the second image does not meet the preset quality conditions, it indicates that the quality of the second image is poor. The computer device can then determine seed points based on the positions of the first and second peaks corresponding to the target peaks and troughs of each parameter curve. For example, the computer device can use the positions of the first and second peaks as seed points; alternatively, it can average the positions of the first and second peaks to obtain the seed points.

[0159] S1302, the third image is obtained by performing flood filling processing based on the seed points.

[0160] In this embodiment, after determining the seed point, the computer device can perform a flood filling process based on the seed point to obtain the third image. It is understood that the third image also includes the connected region where the neck transparency layer is located.

[0161] S1303, if the third image meets the preset quality conditions, the third image is used as the seed image.

[0162] Furthermore, if the third image meets the preset quality conditions, the computer device will use the third image as the seed image.

[0163] In the above implementation, since the second image does not meet the preset quality conditions, the seed point is first determined according to the position of the first peak and the position of the second peak in the target peak and valley corresponding to each parameter curve. Then, the third image is obtained by flood filling based on the seed point. Only when the third image meets the preset quality conditions is the third image used as the seed image, which further improves the accuracy of the seed image.

[0164] Figure 14This is a schematic diagram of another process for determining a seed image in an embodiment of this application. In an exemplary embodiment, such as... Figure 14 As shown, the above parameter determination method also includes S1401 to S1402.

[0165] S1401, if the third image does not meet the preset quality conditions, determine the mask image based on the position of the first peak and the position of the second peak in the alignment of the boundary of the region of interest with the target peaks and valleys corresponding to each parameter curve.

[0166] In this embodiment, if the third image still does not meet the preset quality conditions, it means that the quality of the third image is also poor. The computer device can determine the mask image based on the position of the first peak and the position of the second peak corresponding to the boundary of the region of interest and the target peak and trough of each parameter curve.

[0167] Optionally, the computer device fits the position of the first peak corresponding to the target peak and trough of each parameter curve to obtain the first curve, and fits the position of the second peak corresponding to the target peak and trough of each parameter curve to obtain the second curve. Then, the first and second curves intersect with the boundary of the region of interest, and the computer device retains only the regions corresponding to the four intersection points of the first and second curves with the boundary of the region of interest as the mask image.

[0168] It is understandable that since the mask image is determined based on the positions of the first and second peaks in the target peak-valley pair, the mask image also includes the connected region where the neck transparent layer is located.

[0169] S1402, perform erosion processing on the mask image to obtain the seed image.

[0170] Furthermore, the computer device performs erosion processing on the mask image to obtain a seed image. This erosion processing includes, but is not limited to, morphological erosion processing.

[0171] In the above embodiments, since the third image does not meet the preset quality conditions, the mask image is determined based on the position of the first peak and the position of the second peak corresponding to the boundary of the region of interest and the target peak and valley of each parameter curve, and the mask image is eroded to obtain the seed image. Therefore, the reliability of the obtained seed image is improved.

[0172] In an exemplary embodiment, optionally, the parameter determination method described above further includes the following steps:

[0173] If the aspect ratio of the connected components in the region to be judged is within a preset aspect ratio range, the area of ​​the connected components in the region to be judged is within a preset area range, and the number of pixels belonging to the connected components in the boundary region of the region to be judged is less than a preset number of pixels, then the region to be judged is determined to meet the preset quality conditions; the region to be judged includes the second image or the third image.

[0174] In this embodiment, the preset quality conditions may include a first quality condition, a second quality condition, and a third quality condition.

[0175] The first quality condition is used to determine the aspect ratio of the region to be judged. Since the neck transparent layer is often elongated, if the aspect ratio of the connected regions in the region to be judged is within the preset aspect ratio range, the first quality condition is satisfied; otherwise, it is not satisfied.

[0176] The second quality condition is used to determine the area of ​​the region to be judged. Since the extracted target region has been limited to a relatively small size based on the peaks and troughs in the parameter curves, the area of ​​the connected domains in the region to be judged is relatively fixed relative to the area of ​​the target region. Therefore, if the area of ​​the connected domains in the region to be judged is within the preset area range, the second quality condition is satisfied; otherwise, it is not satisfied.

[0177] The third quality condition is used to determine the adhesion between the region to be judged and its boundary. Since the dark liquid region of the neck transparent layer is always located in the middle, meaning that the connected components in the region to be judged will not adhere much to the upper or lower or left and right boundaries of the region to be judged, the third quality condition is satisfied if the number of pixels belonging to the connected components in the boundary region of the region to be judged is less than the preset number of pixels; otherwise, it is not satisfied. The boundary region of the region to be judged can be determined based on the boundary of the region to be judged, and the size and shape of the boundary region can be set according to requirements.

[0178] It is evident that the first, second, and third quality conditions can determine whether the morphology of the connected domain in the region to be judged conforms to the neck transparency layer from three aspects.

[0179] Furthermore, the computer device will only determine that the region to be judged meets the preset quality conditions if the first, second, and third quality conditions are met. In other words, the computer device determines that the region to be judged meets the preset quality conditions if the aspect ratio of the connected components in the region to be judged is within a preset aspect ratio range, the area of ​​the connected components in the region to be judged is within a preset area range, and the number of pixels belonging to the connected components in the boundary region of the region to be judged is less than a preset number of pixels.

[0180] Alternatively, if at least one of the first quality condition, the second quality condition, and the third quality condition is not met, then the region to be judged is determined to not meet the preset quality condition.

[0181] The region to be judged includes either the second image or the third image. Taking the region to be judged as including the second image as an example, after the computer device processes the first image to obtain the second image, if the aspect ratio of the connected components in the second image is within a preset aspect ratio range, the area of ​​the connected components in the region to be judged is within a preset area range, and the number of pixels belonging to the connected components in the boundary region of the region to be judged is less than a preset number of pixels, then the second image is determined to meet the preset quality conditions, and the second image is used as the seed image.

[0182] In this embodiment, since the region to be judged includes the second image or the third image, if the aspect ratio of the connected domain in the region to be judged is within a preset aspect ratio range, the area of ​​the connected domain in the region to be judged is within a preset area range, and the number of pixels belonging to the connected domain in the boundary region of the region to be judged is less than a preset number of pixels, then the region to be judged is determined to meet the preset quality conditions, which can improve the accuracy of the second image or the third image.

[0183] To more clearly illustrate the process of determining the seed image in this application, this document combines... Figure 15 illustrate. Figure 15 This is a schematic diagram illustrating a process for obtaining a seed image in an embodiment of this application, such as... Figure 15 As shown, after processing the first image to obtain the second image, it is determined whether the second image meets preset quality conditions. If the aspect ratio of the connected components in the second image is within a preset aspect ratio range, the area of ​​the connected components in the region to be judged is within a preset area range, and the number of pixels belonging to the connected components in the boundary region of the region to be judged is less than a preset number of pixels, then the second image is determined to meet the preset quality conditions; otherwise, the second image does not meet the preset quality conditions. Furthermore, if the second image meets the preset quality conditions, it is used as a seed image.

[0184] If the second image does not meet the preset quality conditions, a flood filling process is performed based on the seed points to obtain a third image, and it is then determined whether the third image meets the preset quality conditions. If the aspect ratio of the connected components in the third image is within a preset aspect ratio range, the area of ​​the connected components in the region to be judged is within a preset area range, and the number of pixels belonging to the connected components in the boundary region of the region to be judged is less than a preset number of pixels, then the third image is determined to meet the preset quality conditions; otherwise, the third image does not meet the preset quality conditions. Furthermore, if the third image meets the preset quality conditions, it is used as the seed image.

[0185] If the third image does not meet the preset quality conditions, a mask image is determined, and the mask image is eroded to obtain a seed image. It is evident that this embodiment can handle target regions under different conditions and exhibits strong robustness.

[0186] Figure 16 This is a schematic diagram of another process for determining the contour of the neck transparency layer in an embodiment of this application. In an exemplary embodiment, such as... Figure 16 As shown, S602 also includes S1601 to S1603.

[0187] S1601, Input the seed image into the active contour model to obtain the output result of the current iteration.

[0188] In this embodiment, after the computer device determines the seed image, it inputs the seed image into the active contour model to obtain the current output result. The current output result includes the connected components where the neck transparency layer is located.

[0189] S1602, input the current output result into the active contour model to obtain the next output result.

[0190] The computer then inputs the current output into the active contour model to obtain the next output. The next output includes the connected components containing the neck transparency layer.

[0191] S1603, return to the step of inputting the current output result into the active contour model to obtain the next output result, until the iteration stop condition is met, and determine the contour of the neck transparency layer based on the output result that meets the iteration stop condition; wherein, the iteration stop condition includes the number of iterations reaching a preset number, or the difference between the next output result and the current output result is less than a preset difference.

[0192] Furthermore, the computer device iterates over the seed image as the starting contour of the active contour model. That is, it returns to the step of inputting the current output into the active contour model to obtain the next output, until the iteration stop condition is met.

[0193] For example, the computer device inputs a seed image into the active contour model to obtain the first output result of the active contour model. Then, the computer device inputs the first output result into the active contour model to obtain the second output result of the active contour model. Further, the computer device can continue to input the second output result into the active contour model to obtain the third output result of the active contour model, and so on, until the iteration stopping condition is met.

[0194] The iteration stopping condition includes reaching a preset number of iterations. The preset number can be set according to requirements and is an integer greater than 1. For example, assuming the preset number is 100, the iteration can stop after obtaining the 100th output result of the active contour model.

[0195] Alternatively, the iteration stopping condition includes the following: the difference between the next output and the current output is less than a preset difference. The difference between the next output and the current output is the difference between two adjacent outputs. In other words, each time the computer receives the next output, it can calculate the difference between the next output and the current output, and stop iterating when the difference is less than a preset difference.

[0196] The next output can be represented by the difference between the previous output and the current output, indicating the change in the connected components between them. For example, if the difference between the output of the 50th iteration and the output of the 49th iteration is less than a preset difference, then the iteration can stop.

[0197] Furthermore, the computer device can determine the contour of the neck transparency layer based on the output results that meet the iteration stopping conditions. For example, the computer device can directly use the output results that meet the iteration stopping conditions as the contour of the neck transparency layer, or it can use the output results that meet the iteration stopping conditions after filtering or other processing as the contour of the neck transparency layer.

[0198] Figure 17 This is a schematic diagram of the outline of a neck transparency layer in an embodiment of this application, as shown below. Figure 17 As shown, Figure 17 (a) is a schematic diagram of a seed image. Figure 17 (b) The contour of the neck transparency layer is determined based on the output results that meet the iteration stopping condition after the seed image is input into the contour activity model. It can be seen that the contour of the neck transparency layer can reflect the dark liquid region in the neck transparency layer.

[0199] In the above embodiments, since the iteration stopping condition includes the number of iterations reaching a preset number, or the difference between the next output result and the current output result being less than a preset difference, the seed image is input into the active contour model to obtain the current output result, and the current output result is input into the active contour model to obtain the next output result. Then, the process returns to the step of inputting the current output result into the active contour model to obtain the next output result, until the iteration stopping condition is met. The contour of the neck transparent layer is determined based on the output result that meets the iteration stopping condition. This can improve the accuracy of the obtained neck transparent layer contour during the continuous iteration process, thus playing a role in fine segmentation.

[0200] Figure 18 This is a schematic diagram of another process for determining the contour of the neck transparency layer in an embodiment of this application. In an exemplary embodiment, such as... Figure 18As shown, "determine the contour of the neck transparent layer based on the output result that satisfies the iteration stopping condition" in S1603 also includes S1801 to S1803.

[0201] S1801, determine the intersection-union ratio between the seed image and each candidate connected component in the output result.

[0202] In this embodiment, since the seed image may also contain some impurity regions, the active contour output may also include multiple connected components in the output result that meets the iteration stopping condition. To determine the connected component where the neck transparency layer is located, the computer device determines the intersection-union ratio (IUR) between the seed image and each candidate connected component in the output result that meets the iteration stopping condition. Specifically, determining the IUR between the seed image and each candidate connected component in the output result that meets the iteration stopping condition is equivalent to determining the IUR between the connected component where the neck transparency layer is located in the seed image and each candidate connected component.

[0203] For example, assuming that the output results that satisfy the iteration stopping condition include candidate connected component 1, candidate connected component 2 and candidate connected component 3, the computer device will determine the cross-over ratio 1 between the seed image and candidate connected component 1, the cross-over ratio 2 between the seed image and candidate connected component 2 and the cross-over ratio 3 between the seed image and candidate connected component 3.

[0204] S1802, determine the target connected component based on the candidate connected component with the largest intersection-union ratio.

[0205] Continuing the example above, assuming the intersection-union ratio 3 > intersection-union ratio 1 > intersection-union ratio 2, the computer device can determine the target connected component based on the candidate connected component 3. For example, the computer device can directly use the candidate connected component 3 as the target connected component.

[0206] S1803, Determine the contour of the neck transparency layer based on the target connected region.

[0207] Therefore, in this embodiment, the computer device can determine the contour of the neck transparency layer based on the target connected region. For example, the computer device can determine the thickness of the neck transparency layer by calculating the maximum distance between the upper and lower boundaries in the contour of the neck transparency layer.

[0208] In the above embodiments, since the intersection-union ratio between the seed image and each candidate connected component in the output result is determined, and the target connected component is determined based on the candidate connected component with the largest intersection-union ratio, the contour of the neck transparency layer can be determined more accurately based on the target connected component.

[0209] Figure 19 This is a schematic diagram illustrating another process for determining the thickness of the neck transparency layer in an embodiment of this application. In an exemplary embodiment, such as... Figure 19As shown, S603 also includes S1901 to S1904.

[0210] S1901, determine the midline of the nuchal translucency based on the upper and lower boundaries of the nuchal translucency.

[0211] In this embodiment, the computer device determines the midline of the neck transparency layer based on its upper and lower boundaries. In other words, the computer device can fit the midline of the neck transparency layer based on its upper and lower boundaries. The midline of the neck transparency layer can be a curve.

[0212] S1902, determine the target calculation point from the centerline.

[0213] In this embodiment, the computer device determines the target calculation point from the centerline. Optionally, the computer device can select any point as the target calculation point, or it can select a point from a preset position on the centerline. This embodiment is not limited to this.

[0214] S1903, based on the centerline and the target calculation point, obtain the target straight line corresponding to the target calculation point.

[0215] Furthermore, the computer device can determine nearby calculation points from the centerline based on the target calculation point, and fit the target straight line corresponding to the target calculation point based on the target calculation point and the calculation points near the target calculation point. The calculation points near the target calculation point can be calculation points whose distance from the target calculation point is less than a third preset distance.

[0216] For example, the computer device can determine P calculation points before the target calculation point and Q calculation points after the target calculation point at fixed distances, where the distances between and after the target calculation points represent two extension directions of the median. Then, based on the aforementioned P calculation points, Q calculation points, and the target calculation point, the computer device fits a target straight line corresponding to the target calculation point. Here, P and Q are integers greater than or equal to 1.

[0217] S1904, the thickness of the neck transparent layer is determined based on the distance between the first intersection point and the second intersection point; the first intersection point is the intersection of the normal of the target line and the upper boundary, and the second intersection point is the intersection of the normal and the lower boundary.

[0218] Furthermore, the computer device can determine the normal to the target line, which intersects the upper and lower boundaries of the neck transparency layer's contour at a first intersection point and a second intersection point, respectively. The computer device then determines the distance between the first and second intersection points, and based on this distance, the thickness of the neck transparency layer can be determined. For example, the computer device uses the distance between the first and second intersection points as the thickness of the neck transparency layer.

[0219] In the above embodiments, since the centerline of the neck transparent layer is determined based on the upper boundary and the lower boundary of the neck transparent layer, and the target calculation point is determined from the centerline, the target line corresponding to the target calculation point is obtained based on the centerline and the target calculation point. Furthermore, the first intersection point is the intersection of the normal of the target line and the upper boundary, and the second intersection point is the intersection of the normal and the lower boundary. Therefore, the thickness of the neck transparent layer can be determined based on the distance between the first intersection point and the second intersection point.

[0220] Figure 20 This is a schematic diagram illustrating another process for determining the thickness of the neck transparency layer in an embodiment of this application. In one exemplary embodiment, the number of target calculation points is multiple; such as Figure 20 As shown, S1904 also includes S2001 to S2004.

[0221] S2001, for each target calculation point, the distance between the first intersection point and the second intersection point is taken as the first candidate distance.

[0222] In this embodiment, there are multiple target calculation points. Optionally, the computer device can determine a target calculation point from the centerline at preset length intervals. For each target calculation point, a first intersection point and a second intersection point can be obtained in the manner described above. Then, the distance between the first intersection point and the second intersection point is used as a first candidate distance. For example, assuming there are 5 target calculation points, 5 first candidate distances can be calculated.

[0223] S2002, determine the index value between the first candidate distances corresponding to each target calculation point; the index value includes the standard deviation.

[0224] Furthermore, the computer device can determine the index values ​​between the first candidate distances corresponding to each target calculation point. These index values ​​include the standard deviation, and in some embodiments, they may also include at least one of variance, skewness coefficient, and mean.

[0225] Continuing with the example above, the computer device can calculate the average of the five first candidate distances, and then calculate the standard deviation of the five first candidate distances.

[0226] S2003, based on the index value, the first candidate distance is filtered to obtain the second candidate distance.

[0227] The computer equipment can determine a reference range based on the standard deviation and determine whether the first candidate distance is within the reference range. The first candidate distances outside the reference range are eliminated to obtain the second candidate distance. For example, if there are 5 first candidate distances, namely first candidate distance A to first candidate distance E, after filtering the first candidate distances according to the index value, first candidate distance E is eliminated, and the second candidate distance includes first candidate distances A to first candidate distance D.

[0228] S2004, take the maximum value among the second candidate distances as the thickness of the neck transparency layer.

[0229] The computer device can then use the maximum value among the second candidate distances as the thickness of the neck lumen. Continuing the example above, if the first candidate distance C is the largest among the first candidate distances A to D, then the first candidate distance C is the thickness of the neck lumen.

[0230] In the above embodiments, since the distance between the first intersection point and the second intersection point can be used as the first candidate distance for each target calculation point, and an index value is determined between the first candidate distances corresponding to each target calculation point (including the standard deviation), the first candidate distances can be filtered based on the index value to obtain the second candidate distance, and the maximum value among the second candidate distances can be used as the thickness of the neck transparency layer. Furthermore, since there are multiple target calculation points, the accuracy of the neck transparency layer thickness is further improved.

[0231] Figure 21 This is a schematic diagram of a process for determining a parameter curve in an embodiment of this application. In an exemplary embodiment, such as... Figure 21 As shown, S601 also includes S2101 to S2103.

[0232] S2101, perform a second process on the region of interest to obtain the initial region.

[0233] In this embodiment, after the computer device determines the region of interest (ROI) in the neck translucency layer of the target medical image, it performs a second processing on the ROI to obtain an initial region. This second processing includes, but is not limited to, filtering. The filtering process includes, but is not limited to, at least one of mean filtering, Gaussian filtering, median filtering, or anisotropic diffusion filtering.

[0234] In one embodiment, the initial region exhibits better generalization performance because it can perform secondary processing on the region of interest selected by the doctor, which is beneficial for improving subsequent segmentation accuracy and efficiency.

[0235] For example, a computer device can perform anisotropic diffusion filtering on the region of interest to smooth the image and suppress speckle noise to some extent, thereby improving the accuracy of subsequent target region extraction and threshold segmentation processes and reducing interference in the process of screening connected components.

[0236] S2102, determine multiple target pixel columns in the initial region.

[0237] Furthermore, the computer device can determine multiple target pixel columns from the initial region. It is understood that a target pixel column is a column of pixels within the initial region. Optionally, the computer device can divide the initial region equally to obtain multiple target pixel columns within the region of interest. The number of target pixel columns can be determined as needed. For example, assuming the initial region comprises 180 rows and 200 columns of pixels, the computer device can determine eight target pixel columns within the initial region. The first target pixel column consists of all pixels in the 20th column of the initial region, the second target pixel column consists of all pixels in the 40th column, and so on, with the eighth target pixel column consisting of all pixels in the 180th column. In other words, the computer device can determine one column of pixels as a target pixel column every 20 pixels along the row direction of the initial region.

[0238] S2103, for each target pixel column, determine the parameter curve of the target pixel column based on the grayscale information of the target pixel column.

[0239] In this embodiment, the parametric curve of the target pixel column can be determined based on the grayscale information of each target pixel column. For example, a computer device fits each pixel in the target pixel column sequentially along the column direction to obtain the parametric curve of the target pixel column.

[0240] In the above embodiments, since the region of interest can be processed a second time to obtain the initial region, the accuracy of the initial region is improved. Furthermore, since multiple target pixel columns can be determined in the initial region, and for each target pixel column, the parameter curve of the target pixel column can be determined based on the grayscale information of the target pixel column, that is, multiple parameter curves of the region of interest are obtained.

[0241] Figure 22 This is a schematic diagram of a process for determining a target pixel column according to an embodiment of this application. In an exemplary embodiment, such as... Figure 22 As shown, S2102 also includes S2201 to S2203.

[0242] S2201, the initial region is divided to obtain multiple first candidate columns.

[0243] In this embodiment, the computer device divides the initial region to obtain multiple first candidate columns. Optionally, the computer device can divide the initial region equally to obtain multiple first candidate columns. For example, the computer device can divide the region of interest into 9 equal parts in the vertical direction and obtain 8 first candidate columns.

[0244] S2202, determine the second candidate column from the first candidate column based on the position of the first candidate column in the initial region.

[0245] Furthermore, after determining the first candidate column, a second candidate column can be determined from the first candidate column based on its position within the initial region. Optionally, the computer device can use the first candidate column located at a preset position within the initial region as the second candidate column. For example, if the computer device uses the first candidate column located in the middle as the second candidate column, and assuming the eight first candidate columns are first candidate column 1, first candidate column 2...first candidate column 8, then the computer device can use first candidate columns 3 to 5 as the second candidate columns.

[0246] S2203, based on the size of the target medical image, expand the pixel range of the second candidate column to obtain the target pixel column.

[0247] Optionally, the computer device can expand the pixel range of the second candidate column to the left and / or right to obtain the target pixel column based on the size of the target medical image. For example, the computer device can expand the second candidate column to the left or right by a preset number of pixels, where the preset number of pixels is 1 / 10 of the width of the target medical image.

[0248] Alternatively, since the target pixel column has a certain width, the computer device can sum or average the target pixel column horizontally and then fit it vertically to obtain the parameter curve.

[0249] In the above embodiment, the initial region is first divided to obtain multiple first candidate columns. Then, based on the position of the first candidate columns within the initial region, second candidate columns are determined from the first candidate columns. Finally, based on the size of the target medical image, the pixel range of the second candidate columns is expanded to obtain the target pixel column. In this way, multiple target pixel columns are determined from the initial region.

[0250] Figure 23 This is a visual flowchart illustrating an embodiment of the present application. In one exemplary embodiment, such as... Figure 23 As shown, the above parameter determination method also includes S2301 to S2303.

[0251] S2301, Determine the size information of the target medical image.

[0252] In this embodiment, the size information can be the size obtained by the computer device after receiving the target medical image and analyzing it, or it can be the size determined in response to the doctor's input operation. The size information is used to characterize the physical size or actual size of the target medical image.

[0253] S2302, magnifies the area of ​​interest by a preset factor based on the size information.

[0254] The computer device can store the correspondence between different size information and preset magnification, and then determine the preset magnification of the region of interest based on the size information of the target medical image.

[0255] For example, assuming the size information is magnified 1.5 times in the preset reference range 1 and 2 times in the preset reference range 2, when the size information of the target medical image is in the preset reference range 2, the computer device can magnify the region of interest by 2 times so that the doctor can clearly observe the edge position of the neck translucency layer.

[0256] S2303 displays the magnified region of interest and the thickness of the neck transparent layer in the visualization interface.

[0257] Furthermore, the computer device can provide a visualization interface that displays a magnified view of the region of interest and the thickness of the neck transparency layer.

[0258] In some embodiments, the computer device may also display the target medical image in a visual interface.

[0259] In some embodiments, the computer device may also respond to a doctor's zoom-in or zoom-out operation to adjust the scaling of the region of interest in the visualization interface.

[0260] In some embodiments, the computer device may also display parameters such as the average, variance, and minimum thickness of the neck transparency layer.

[0261] In some embodiments, the computer device may also display the outline of the neck transparency layer.

[0262] Figure 24 This is a schematic diagram of a visual interface in an embodiment of this application. Figure 24 (a) shows the target medical image and the magnified region of interest. The region of interest outlined by the solid white line is magnified to obtain the magnified region of interest shown by the dashed white line. Figure 24 (b) The outline and thickness of the neck transparency layer are shown. Specifically, after analyzing the region of interest outlined by the white solid line, the outline and thickness of the neck transparency layer, for example, 0.12 cm, can be displayed within the white dashed line.

[0263] In the above embodiments, since the size information of the target medical image can be determined, and the region of interest is magnified by a preset factor when the size information is within a preset reference range, the magnified region of interest and the thickness of the neck transparent layer are displayed in the visualization interface, thus achieving adaptive magnification and reducing the doctor's operation process.

[0264] To more clearly illustrate the parameter determination method in the embodiments of this application, this document combines... Figure 25 and Figure 26 Please provide an explanation. Figure 25 This is a flowchart illustrating a parameter determination method according to an embodiment of this application. Figure 25 As shown, the parameter determination method can be divided into three stages: target region extraction, NT region segmentation, and contour extraction and adjustment.

[0265] Figure 26 and Figure 27 This is a schematic diagram illustrating a parameter determination method according to an embodiment of this application. Figure 26 and Figure 27 As shown, the computer device can execute this parameter determination method according to the following procedure.

[0266] S2601, a second processing step is performed on the region of interest in the neck transparency layer of the target medical image to obtain the initial region. S2601 thus achieves... Figure 25 The process of obtaining the region of interest and the initial region.

[0267] S2602, the initial region is divided to obtain multiple first candidate columns.

[0268] S2603, determine the second candidate column from the first candidate column based on the position of the first candidate column in the initial region.

[0269] S2604, based on the size of the target medical image, expand the pixel range of the second candidate column to obtain the target pixel column.

[0270] S2605: For each target pixel column, determine the parameter curve of the target pixel column based on the grayscale information of the target pixel column. That is, obtain multiple parameter curves for the region of interest.

[0271] S2606, Based on the peaks and troughs in each parameter curve, extract the peak-trough pairs from each parameter curve. The peak-trough pair includes the first peak, the second peak, and the trough located between the first peak and the second peak.

[0272] S2607 uses the parameter curve with the largest number of peak-to-trough pairs as the reference parameter curve.

[0273] S2608, for each reference peak-valley pair on the reference parameter curve, determine the distance between the reference peak-valley pair and peak-valley pairs on other parameter curves. The other parameter curves are parameter curves other than the reference parameter curve.

[0274] S2609, determine the candidate groups corresponding to each reference peak-valley pair based on the distance. The candidate groups include peak-valley pairs located in each parameter curve.

[0275] S2610, for each candidate group, determine the first ratio between the summation result and the maximum peak-to-trough ratio. The summation result is the sum of the peak-to-trough ratios of each peak-to-trough pair in the candidate group.

[0276] S2611, for each candidate group, determine the second ratio between the standard deviation of the distance and the maximum standard deviation of the distance. The standard deviation of the distance is determined based on the Euclidean distance between each peak-trough pair in the candidate group.

[0277] S2612, determine the identification score corresponding to the candidate group based on the first ratio and the second ratio.

[0278] S2613, determine the target peak-valley pairs corresponding to each parameter curve based on the discrimination score.

[0279] S2614, the first curve is obtained by fitting the position of the first peak in the target peak-valley pair corresponding to each parameter curve.

[0280] S2615, the second curve is obtained by fitting the position of the second peak in the target peak-valley pair corresponding to each parameter curve.

[0281] S2616: The target region is determined based on the intersection points of the first and second curves with the boundary of the region of interest. S2602 to S2616 thus achieve... Figure 25 The process of extracting the target region.

[0282] S2617, the target region is segmented according to a preset threshold segmentation algorithm to obtain the first image. S2617 thus achieves... Figure 25 The process of segmenting the target region.

[0283] S2618, perform a first processing on the first image to obtain a second image. The first processing includes at least one of morphological erosion, area filtering, and centroid location filtering. Area filtering includes retaining connected components in the first image with an area greater than a preset area; centroid location filtering includes retaining connected components in the first image whose centroids are within a preset centroid range.

[0284] S2619, determine whether the second image meets the preset quality conditions. If it does, proceed to S2620; otherwise, proceed to S2621.

[0285] S2620: If the second image meets the preset quality conditions, the second image is used as the seed image. Then proceed to S2627.

[0286] S2621, if the second image does not meet the preset quality conditions, determine the seed point according to the position of the first peak and the position of the second peak in the target peak-valley alignment corresponding to each parameter curve.

[0287] S2622, the third image is obtained by performing flood filling processing based on the seed points.

[0288] S2623, Determine whether the third image meets the preset quality conditions. If it does, proceed to S2624; otherwise, proceed to S2625.

[0289] S2624: If the third image meets the preset quality conditions, the third image is used as the seed image. Then proceed to S2627.

[0290] S2625, if the third image does not meet the preset quality conditions, the mask image is determined based on the alignment of the boundary of the region of interest with the positions of the first and second peaks of the target peaks and troughs corresponding to each parameter curve. S2618 to S2625 thus achieve... Figure 25 The process of morphological and connected component analysis in China.

[0291] S2626: Erosion processing is performed on the mask image to obtain a seed image. Then, proceed to S2627. Specifically, if the aspect ratio of the connected components in the region to be judged is within a preset aspect ratio range, the area of ​​the connected components in the region to be judged is within a preset area range, and the number of pixels belonging to the connected components in the boundary region of the region to be judged is less than a preset number of pixels, then the region to be judged is determined to meet the preset quality conditions; the region to be judged includes either the second image or the third image. S2626 thus achieves... Figure 25 The process of obtaining a seed image.

[0292] S2627, Input the seed image into the active contour model to obtain the output result of the current iteration.

[0293] S2628: Input the current output result into the active contour model to obtain the next output result.

[0294] S2629 returns to the step of inputting the current output result into the active contour model to obtain the next output result, until the iteration stopping condition is met. The iteration stopping condition includes either reaching a preset number of iterations, or the difference between the next output result and the current output result being less than a preset difference. S2627–S2629 thus achieves this. Figure 25The process of iterating the active contour model.

[0295] S2630, determine the intersection-union ratio between the seed image and each candidate connected component in the output result that satisfies the iteration stopping condition.

[0296] S2631 determines the target connected component based on the candidate connected component with the largest intersection-union ratio. S2630 to S2631 thus achieve... Figure 25 The process of analyzing connected components in the middle.

[0297] S2632, determines the contour of the neck transparency layer based on the target connected components. S2632 thus achieves... Figure 25 The process of adjusting the mid-contour and extracting the top and bottom edges.

[0298] S2633, determine the midline of the nuchal translucency based on the upper and lower boundaries of the nuchal translucency.

[0299] S2634, determine the target calculation points from the centerline. There are multiple target calculation points.

[0300] S2635, based on the centerline and the target calculation point, obtain the target straight line corresponding to the target calculation point.

[0301] S2636, for each target calculation point, the distance between the first intersection point and the second intersection point is taken as the first candidate distance.

[0302] S2637, determine the index values ​​between the first candidate distances corresponding to each target calculation point. The index values ​​include the standard deviation.

[0303] S2638, based on the index value, the first candidate distance is filtered to obtain the second candidate distance.

[0304] S2639, the maximum value among the second candidate distances is used as the thickness of the neck transparency layer. S2639 thus achieves... Figure 25 The process of determining the thickness of the neck transparency layer.

[0305] S2640, determine the size information of the target medical image.

[0306] S2641, magnifies the region of interest by a preset factor based on the size information. S2641 thus achieves... Figure 25 The process of scaling the region of interest.

[0307] S2642 displays the magnified region of interest and the thickness of the neck transparency layer in the visualization interface.

[0308] S2601 to S1642 can be referred to the above embodiments and will not be repeated here. It is evident that traditional parameter determination methods, on the one hand, rely on the doctor's experience, resulting in subjective measurement results and low accuracy; on the other hand, since doctors typically only measure the maximum thickness of the nuchal translucency, the measurement results have significant uncertainty and bias. In this embodiment, on the one hand, by extracting multiple distances from the contour of the nuchal translucency and then selecting the maximum distance as the nuchal translucency thickness, the accuracy of the measurement results is improved. On the other hand, this embodiment can measure multiple indicators such as the average thickness and standard deviation of the nuchal translucency at once, which is beneficial for the doctor's subsequent evaluation. Furthermore, this embodiment can complete automatic detection using traditional image processing algorithms, requiring only a small number of target medical images for verification, which distinguishes it from existing deep learning methods and improves efficiency.

[0309] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0310] In one embodiment, a parameter determination system is provided. Figure 28 This is a schematic diagram of the architecture of a parameter determination system according to an embodiment of this application, such as... Figure 28 As shown, this parameter determines that the system includes a front-end, a mid-end, and a back-end. The front-end includes an ultrasound transducer and a data receiving module. The ultrasound transducer transmits and receives ultrasound signals; the data receiving module receives, amplifies, and performs analog-to-digital conversion on the electrical signals from the ultrasound transducer, then sends the processed echo digital signals to the mid-end for further processing. The mid-end processing analyzes, interpolates, and filters the echo digital signals to generate a medical image with relatively coarse image quality. Mid-end processing includes logarithmic compression, spatial filtering, gain compensation, and coordinate transformation.

[0311] The medical images generated at the terminal are sent to the backend for image post-processing to obtain the target medical image, such as smoothing and edge enhancement to obtain an ultrasound image. Afterward, the backend stores the target medical image as a data storage device.

[0312] Furthermore, through the parameter determination process of this embodiment, the outline and thickness of the cervical translucency layer can be obtained from the target medical image, and finally displayed in the visualization interface.

[0313] Based on the same inventive concept, this application also provides a parameter determination apparatus for implementing the parameter determination method described above. The solution provided by this apparatus is similar to the implementation scheme described in the above method; therefore, the specific limitations in one or more parameter determination apparatus embodiments provided below can be found in the limitations of the parameter determination method described above, and will not be repeated here.

[0314] Figure 29 This is a structural block diagram of the parameter determination and adjustment device in an embodiment of this application. In one exemplary embodiment, such as... Figure 29 As shown, a parameter determination device 2900 is provided, including: a first determination module 2901, an extraction module 2902, and a second determination module 2903, wherein:

[0315] The first determining module 2901 is used to obtain multiple parameter curves of the region of interest based on the region of interest in the neck transparency layer of the target medical image.

[0316] Extraction module 2902 is used to extract the target region of interest based on the peaks and troughs in the curves of each parameter.

[0317] The second determining module 2903 is used to determine the thickness of the neck transparent layer based on the target area.

[0318] The aforementioned parameter determination device can obtain multiple parametric curves for the region of interest (ROI) of the neck translucency layer in a target medical image. Based on the peaks and troughs of each parametric curve, a target region within the ROI is extracted to determine the thickness of the neck translucency layer. This reduces reliance on physicians in the thickness determination process, thereby minimizing the influence of physician experience on the thickness determination and preventing significant discrepancies between the thicknesses determined by different physicians for the same target medical image, thus improving the consistency of the determined neck translucency layer thickness. Furthermore, since the target region is extracted from the ROI of the neck translucency layer based on the peaks and troughs of each parametric curve, the characteristics of the neck translucency layer on the parametric curves are taken into account during the target region determination process. This narrows the target region, increasing the processing speed based on the target region and thus improving the efficiency of determining the neck translucency layer thickness. On the other hand, it improves the precision of the target region, thereby enhancing the accuracy of the neck translucency layer thickness determined based on the target region.

[0319] Optionally, the second determining module 2903 includes:

[0320] The segmentation unit is used to segment the target region to obtain a seed image.

[0321] The first determining unit is used to determine the contour of the neck transparency layer based on the seed image and the active contour model.

[0322] The second determining unit is used to determine the thickness of the neck transparency layer based on its contour.

[0323] Optionally, the extraction module 2902 includes:

[0324] The extraction unit is used to extract peak-valley pairs from each parameter curve based on the peaks and valleys in each parameter curve; the peak-valley pairs include a first peak, a second peak, and a valley located between the first peak and the second peak.

[0325] The third determining unit is used to determine the discrimination score of peak-valley pairs in each parameter curve.

[0326] The fourth determining unit is used to determine the target peak-valley pairs corresponding to each parameter curve based on the discrimination score.

[0327] The fifth determining unit is used to determine the target area based on the position of the first peak and the position of the second peak in the target peak-valley alignment corresponding to each parameter curve.

[0328] Optionally, the third determining unit includes:

[0329] The first defined sub-unit is used to take the parameter curve with the largest number of peak-valley pairs as the reference parameter curve.

[0330] The second determining sub-unit is used to determine the distance between each reference peak-valley pair on the reference parameter curve and peak-valley pairs on other parameter curves; the other parameter curves are parameter curves other than the reference parameter curve.

[0331] The third determining sub-unit is used to determine the candidate group corresponding to each reference peak-valley pair based on the distance; the candidate group includes peak-valley pairs located in each parameter curve.

[0332] The fourth determination subunit is used to determine the discrimination score corresponding to each candidate group.

[0333] Optionally, the fourth determining subunit is also used to determine, for each candidate group, a first ratio between the summation result and the maximum peak-to-trough ratio; the summation result is the sum of the peak-to-trough ratios of each peak-to-trough pair in the candidate group; for each candidate group, a second ratio between the distance standard deviation and the maximum distance standard deviation; the distance standard deviation is determined based on the Euclidean distance of each peak-to-trough pair in the candidate group; and a discrimination score is determined based on the first ratio and the second ratio.

[0334] Optionally, the fifth determining unit includes:

[0335] The first fitting subunit is used to fit the position of the first peak in the target peak-valley pair corresponding to each parameter curve, and obtain the first curve.

[0336] The second fitting subunit is used to fit the position of the second peak in the target peak-valley pair corresponding to each parameter curve, and obtain the second curve.

[0337] The fifth determining sub-unit is used to determine the target region based on the intersection points of the first curve and the second curve with the boundary of the region of interest, respectively.

[0338] Optionally, the segmentation unit includes:

[0339] The segmentation subunit is used to segment the target region according to a preset threshold segmentation algorithm to obtain the first image.

[0340] The processing subunit is used to perform a first process on the first image to obtain a second image.

[0341] The sixth determining subunit is used to use the second image as a seed image if the second image meets the preset quality conditions.

[0342] Optionally, the first processing includes at least one of morphological erosion processing, area filtering, and centroid location filtering; area filtering includes retaining connected components in the first image with an area greater than a preset area; centroid location filtering includes retaining connected components in the first image whose centroids are within a preset centroid range.

[0343] Optionally, the parameter determining device 2900 further includes:

[0344] The third determining module is used to determine the seed point based on the position of the first peak and the position of the second peak in the alignment of the target peak and trough corresponding to each parameter curve when the second image does not meet the preset quality conditions.

[0345] The first processing module is used to perform flood filling processing based on seed points to obtain the third image.

[0346] The fourth determining module is used to use the third image as a seed image if the third image meets the preset quality conditions.

[0347] Optionally, the parameter determining device 2900 further includes:

[0348] The fifth determining module is used to determine the mask image based on the position of the first peak and the position of the second peak when the third image does not meet the preset quality conditions, according to the boundary of the region of interest and the target peak and valley corresponding to each parameter curve.

[0349] The second processing module is used to perform erosion processing on the mask image to obtain the seed image.

[0350] Optionally, the parameter determining device 2900 further includes:

[0351] The sixth determining module is used to determine that the region to be judged meets the preset quality conditions if the aspect ratio of the connected domain in the region to be judged is within a preset aspect ratio range, the area of ​​the connected domain in the region to be judged is within a preset area range, and the number of pixels belonging to the connected domain in the boundary region of the region to be judged is less than a preset number of pixels; the region to be judged includes the second image or the third image.

[0352] Optionally, the first determining unit includes:

[0353] The sixth determination sub-unit is used to input the seed image into the active contour model to obtain the output result of the current iteration.

[0354] The seventh determination sub-unit is used to input the current output result into the active contour model to obtain the next output result.

[0355] The eighth step is to determine the sub-unit, return to the step of inputting the current output result into the active contour model to obtain the next output result, until the iteration stop condition is met, and determine the contour of the neck transparent layer based on the output result that meets the iteration stop condition; wherein, the iteration stop condition includes the number of iterations reaching a preset number, or the difference between the next output result and the current output result being less than a preset difference.

[0356] Optionally, the eighth determining subunit is also used to determine the intersection-union ratio (IUR) between the seed image and each candidate connected component in the output result; determine the target connected component based on the candidate connected component with the largest IUR; and determine the contour of the neck transparency layer based on the target connected component.

[0357] Optionally, the second determining module 2903 includes:

[0358] The sixth determining unit is used to determine the midline of the neck transparency based on the upper boundary and the lower boundary of the neck transparency.

[0359] The seventh determining unit is used to determine the target calculation point from the centerline.

[0360] The eighth determining unit is used to obtain the target line corresponding to the target calculation point based on the centerline and the target calculation point.

[0361] The ninth determining unit is used to determine the thickness of the neck transparent layer based on the distance between the first intersection point and the second intersection point; the first intersection point is the intersection of the normal of the target line and the upper boundary, and the second intersection point is the intersection of the normal and the lower boundary.

[0362] Optionally, the number of target calculation points can be multiple; the ninth determining unit includes:

[0363] The ninth determining sub-unit is used to determine the distance between the first intersection point and the second intersection point as the first candidate distance for each target calculation point.

[0364] The tenth determination sub-unit is used to determine the index value between the first candidate distances corresponding to each target calculation point; the index value includes the standard deviation.

[0365] The eleventh sub-unit is used to filter the first candidate distance based on the index value to obtain the second candidate distance.

[0366] The twelfth determining sub-unit is used to take the maximum value among the second candidate distances as the thickness of the neck transparency layer.

[0367] Optionally, the first determining module 2901 includes:

[0368] The processing unit is used to perform a second processing on the region of interest to obtain the initial region.

[0369] The tenth determining unit is used to determine multiple target pixel columns in the initial region.

[0370] The eleventh determining unit is used to determine the parameter curve of each target pixel column based on the grayscale information of the target pixel column.

[0371] Optionally, the tenth determining unit includes:

[0372] Divide the initial region into sub-units to obtain multiple first candidate columns.

[0373] The thirteenth determining sub-unit is used to determine the second candidate column from the first candidate column based on the position of the first candidate column in the initial region.

[0374] The expanded sub-unit is used to expand the pixel range of the second candidate column to obtain the target pixel column based on the size of the target medical image.

[0375] Optionally, the parameter determining device 2900 further includes:

[0376] The twelfth determining unit is used to determine the size information of the target medical image.

[0377] The magnification unit is used to magnify the area of ​​interest by a preset factor based on the size information.

[0378] The display unit is used to show the magnified region of interest and the thickness of the neck transparent layer in the visualization interface.

[0379] The modules in the aforementioned parameter determining device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0380] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0381] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0382] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0383] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0384] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0385] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for determining parameters, characterized in that, The method includes: Based on the region of interest in the transparent layer of the neck in the target medical image, multiple parametric curves of the region of interest are obtained, including grayscale curves, gradient curves and pixel value curves. Based on the peaks and troughs in the parameter curves, extract the target region from the region of interest. The thickness of the neck transparent layer is determined based on the target region; The step of extracting the target region from the region of interest based on the peaks and troughs in each of the parameter curves includes: Based on the peaks and troughs in each parameter curve, peak-trough pairs are extracted from each parameter curve; the peak-trough pairs include a first peak, a second peak, and a trough located between the first peak and the second peak; Determine the discrimination score of peak-valley pairs in each of the parameter curves; The target peak-valley pairs corresponding to each parameter curve are determined based on the discrimination score. The first curve is obtained by fitting the position of the first peak in the target peak-valley pair corresponding to each of the parameter curves; the second curve is obtained by fitting the position of the second peak in the target peak-valley pair corresponding to each of the parameter curves; the target region is determined based on the intersection points of the first curve and the second curve with the boundary of the region of interest.

2. The method according to claim 1, characterized in that, Determining the thickness of the neck transparency layer based on the target region includes: The target region is segmented to obtain a seed image; The contour of the neck transparency layer is determined based on the seed image and the active contour model; The thickness of the neck transparency layer is determined based on its contour.

3. The method according to claim 1, characterized in that, The determination of the discrimination score of peak-valley pairs in each of the parameter curves includes: The parameter curve with the largest number of peak-valley pairs is used as the reference parameter curve. For each reference peak-valley pair on the reference parameter curve, determine the distance between the reference peak-valley pair and peak-valley pairs on other parameter curves; the other parameter curves are parameter curves other than the reference parameter curve. Based on the distance, candidate groups corresponding to each of the reference peak-valley pairs are determined; the candidate groups include peak-valley pairs located in each of the parameter curves. Determine the discrimination score corresponding to each candidate group.

4. The method according to claim 3, characterized in that, Determining the discrimination score corresponding to each candidate group includes: For each candidate group, a first ratio is determined between the summation result and the maximum peak-to-trough ratio; the summation result is the sum of the peak-to-trough ratios of each peak-to-trough pair in the candidate group; For each candidate group, a second ratio between the distance standard deviation and the maximum distance standard deviation is determined; the distance standard deviation is determined based on the Euclidean distance between each peak-trough pair in the candidate group. The discrimination score is determined based on the first ratio and the second ratio.

5. The method according to claim 2, characterized in that, The step of segmenting the target region to obtain a seed image includes: The target region is segmented according to a preset threshold segmentation algorithm to obtain a first image; The first image is processed to obtain the second image; If the second image meets the preset quality conditions, the second image is used as the seed image.

6. The method according to claim 5, characterized in that, The first process includes at least one of morphological etching, area screening, and centroid location screening; The area filtering includes retaining connected components in the first image whose area is greater than a preset area; The centroid location filtering includes retaining connected components in the first image whose centroids are within a preset centroid range.

7. The method according to claim 6, characterized in that, The method further includes: If the second image does not meet the preset quality conditions, seed points are determined based on the position of the first peak and the position of the second peak in the target peak-valley alignment corresponding to each parameter curve. The third image is obtained by performing a flood filling process based on the seed points; If the third image meets the preset quality conditions, the third image is used as the seed image.

8. The method according to claim 7, characterized in that, The method further includes: If the third image does not meet the preset quality conditions, the mask image is determined based on the position of the first peak and the position of the second peak in the alignment of the boundary of the region of interest with the target peaks and valleys corresponding to each parameter curve. The seed image is obtained by erosion processing the mask image.

9. The method according to claim 8, characterized in that, The method further includes: If the aspect ratio of the connected components in the region to be judged is within a preset aspect ratio range, the area of ​​the connected components in the region to be judged is within a preset area range, and the number of pixels belonging to the connected components in the boundary region of the region to be judged is less than a preset number of pixels, then the region to be judged is determined to meet the preset quality condition. The region to be judged includes either the second image or the third image.

10. The method according to claim 2, characterized in that, Determining the contour of the neck transparency layer based on the seed image and the active contour model includes: The seed image is input into the active contour model to obtain the output result for the current iteration; The current output result is input into the active contour model to obtain the next output result; Return to the step of inputting the current output result into the active contour model to obtain the next output result, until the iteration stop condition is met, and determine the contour of the neck transparency layer based on the output result that meets the iteration stop condition; The iteration stopping conditions include the number of iterations reaching a preset number, or the difference between the next output result and the current output result being less than a preset difference.

11. The method according to claim 10, characterized in that, Determining the contour of the neck transparency layer based on the output result satisfying the iteration stopping condition includes: Determine the intersection-union ratio (IUU) between the seed image and each candidate connected component in the output result; The target connected component is determined based on the candidate connected component with the largest intersection-union ratio. The contour of the neck transparency layer is determined based on the target connected region.

12. The method according to any one of claims 1-4, characterized in that, Determining the thickness of the neck transparency layer based on the target region includes: The midline of the neck transparent layer is determined based on the upper boundary and the lower boundary of the neck transparent layer; Determine the target calculation point from the centerline; Based on the centerline and the target calculation point, the target straight line corresponding to the target calculation point is obtained; The thickness of the neck transparent layer is determined based on the distance between the first intersection point and the second intersection point; the first intersection point is the intersection of the normal of the target line and the upper boundary, and the second intersection point is the intersection of the normal and the lower boundary.

13. The method according to claim 12, characterized in that, The number of target calculation points is multiple; determining the thickness of the neck transparency layer based on the distance between the first intersection point and the second intersection point includes: For each of the target calculation points, the distance between the first intersection point and the second intersection point is taken as the first candidate distance; Determine the index values ​​between the first candidate distances corresponding to each of the target calculation points; the index values ​​include the standard deviation. The first candidate distance is filtered based on the index value to obtain the second candidate distance; The maximum value among the second candidate distances is taken as the thickness of the neck transparency layer.

14. The method according to any one of claims 1-4, characterized in that, The region of interest, based on the neck transparency layer in the target medical image, is used to obtain multiple parameter curves for the region of interest, including: The region of interest is processed a second time to obtain the initial region; Determine multiple target pixel columns in the initial region; For each target pixel column, the parameter curve of the target pixel column is determined based on the grayscale information of the target pixel column.

15. The method according to claim 14, characterized in that, Determining the multiple target pixel columns in the initial region includes: The initial region is divided to obtain multiple first candidate columns; Based on the position of the first candidate column in the initial region, a second candidate column is determined from the first candidate column; The target pixel column is obtained by expanding the pixel range of the second candidate column based on the size of the target medical image.

16. The method according to any one of claims 1-4, characterized in that, The method further includes: Determine the size information of the target medical image; The region of interest is magnified by a preset factor based on the size information; The magnified region of interest and the thickness of the transparent neck layer are displayed in the visualization interface.

17. A parameter determining device, characterized in that, The device includes: The first determining module is used to obtain multiple parametric curves of the region of interest based on the region of interest in the transparent layer of the neck in the target medical image. The multiple parametric curves include grayscale curves, gradient curves, and pixel value curves. The extraction module is used to extract the target region from the region of interest based on the peaks and troughs in the parameter curves. The second determining module is used to determine the thickness of the neck transparent layer based on the target region; The step of extracting the target region from the region of interest based on the peaks and troughs in each of the parameter curves includes: Based on the peaks and troughs in each parameter curve, peak-trough pairs are extracted from each parameter curve; the peak-trough pairs include a first peak, a second peak, and a trough located between the first peak and the second peak; Determine the discrimination score of peak-valley pairs in each of the parameter curves; The target peak-valley pairs corresponding to each parameter curve are determined based on the discrimination score. The first curve is obtained by fitting the position of the first peak in the target peak-valley pair corresponding to each of the parameter curves; the second curve is obtained by fitting the position of the second peak in the target peak-valley pair corresponding to each of the parameter curves; the target region is determined based on the intersection points of the first curve and the second curve with the boundary of the region of interest.

18. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 16.

Images