A method and system for assisting in recognizing nursing images in reproductive medicine

Abstract

The present invention relates to the field of image data processing technology, and specifically to a reproductive medicine nursing image auxiliary recognition method and system. The present invention obtains the connected domain of the follicle image; according to the edge morphological characteristics of each connected domain, the possibility of the suspected follicle area of ​​each connected domain is obtained, and the suspected follicle area of ​​each frame of the follicle image is screened out; the position distribution and morphological characteristics of the suspected follicle area between different frames of follicle images are analyzed to obtain multiple groups of matching suspected follicle areas, and the possibility of the same follicle in each group of matching suspected follicle areas is obtained; the same follicle marking area on different frames of follicle images is screened out; the present invention improves the analysis accuracy of follicle development by accurately identifying the same follicle area corresponding to different frames of follicle images.

Description

Technical Field

[0001] The present invention relates to the technical field of image data processing, and in particular to a method and system for assisting in recognizing nursing images in reproductive medicine. Background Art

[0002] During assisted reproduction, the time when the ovarian follicles in a woman's body mature is particularly important. Accurately identifying the changes in the ovarian follicles can improve the success rate of assisted reproduction. Therefore, reproductive medicine nursing usually uses transvaginal ultrasound technology to obtain follicle images and further identify the follicles.

[0003] In the prior art, edge detection is used to identify follicles in the acquired images, that is, to find closed edge curves and locate the follicle area with a fixed feature template; however, since there are multiple follicles in the image and the morphological characteristics of follicles at different developmental stages are also different, using only a certain fixed template for identification may result in misjudgment and missed judgment, inaccurate determination of the developmental follicle area, and poor follicle identification effect. Summary of the invention

[0004] In order to solve the technical problems of inaccurate determination of the developing follicle area and poor follicle recognition effect, the purpose of the present invention is to provide a reproductive medicine nursing image-assisted recognition method and system, and the technical scheme adopted is as follows:

[0005] The present invention proposes a reproductive medicine nursing image-assisted recognition method, the method comprising:

[0006] Acquire multiple frames of follicle images of patients preparing for pregnancy in time sequence;

[0007] For any frame of follicle image, the connected domain of the follicle image is obtained according to the grayscale distribution of the pixels on the follicle image; the possibility of the suspected follicle region of each connected domain is obtained according to the edge morphological characteristics of each connected domain; and the suspected follicle region of each frame of follicle image is screened out according to the possibility of the suspected follicle region;

[0008] According to the position distribution and morphological characteristics of the suspected follicle regions between different frames of follicle images, multiple groups of matching suspected follicle regions are obtained; according to the position change characteristics and morphological change characteristics of each group of matching suspected follicle regions, the possibility of the same follicle in each group of matching suspected follicle regions is obtained;

[0009] According to the same follicle possibility, the same follicle marked area on the follicle images of different frames is screened out.

[0010] Furthermore, the method for obtaining the possibility of the suspected follicle area includes:

[0011] For each connected domain, obtain the difference between the 8-chain code value of each edge pixel and the previous adjacent edge pixel as the chain code difference;

[0012] According to the absolute value of the average chain code difference between all edge pixels and the previous adjacent edge pixels, as well as the fluctuation characteristics of the chain code difference, the possibility of the suspected follicle area of ​​each connected domain is obtained. The absolute value of the average chain code difference and the fluctuation characteristics of the chain code difference are negatively correlated with the possibility of the suspected follicle area.

[0013] Furthermore, the method for obtaining the suspected follicle region includes:

[0014] For any frame of follicle image, if the probability of the suspected follicle region in the connected domain is greater than or equal to the preset probability threshold, the corresponding connected domain is taken as the suspected follicle region.

[0015] Furthermore, the method for obtaining the matching suspected follicle region includes:

[0016] Obtain the centroid coordinates of each suspected follicle region in each frame of follicle image;

[0017] In each frame of follicle image, a neighborhood range is constructed with the centroid coordinates of each suspected follicle region of the previous frame of follicle image as the center. If the number of pixel points in the suspected follicle region in the corresponding neighborhood range is greater than the number of pixel points in the suspected follicle region corresponding to the previous frame of follicle image, the corresponding suspected follicle region in the neighborhood range is used as a candidate suspected follicle region; the candidate suspected follicle region with the smallest relative distance between the corresponding centroid coordinates and the center position of the neighborhood range is selected as the same type of suspected follicle region corresponding to the suspected follicle region of the previous frame of follicle image;

[0018] If each suspected follicle region has similar suspected follicle regions in different frames of follicle images, a group of matching suspected follicle regions is formed.

[0019] Furthermore, the method for obtaining the possibility of the same follicle includes:

[0020] For any group of matched suspected follicle regions, the ratio of the number of pixels in the matched suspected follicle region between each frame of follicle image and the previous frame of follicle image is obtained as the morphological enlargement rate of the matched suspected follicle region corresponding to each frame of follicle image;

[0021] Obtaining the difference in centroid coordinates in the matching suspected follicle region between each frame of follicle image and the previous frame of follicle image as a displacement vector of the matching suspected follicle region corresponding to each frame of follicle image;

[0022] According to the difference in morphological enlargement rate and displacement vector of the corresponding matching suspected follicle area between adjacent frames of follicle images, the possibility of the same follicle in the corresponding group of matching suspected follicle areas is obtained. The difference in morphological enlargement rate and displacement vector are negatively correlated with the possibility of the same follicle.

[0023] Furthermore, the method for obtaining the same follicle marked area includes:

[0024] If the probability of a group of matching suspected follicle regions being the same follicle is greater than or equal to a preset same follicle threshold, the corresponding matching suspected follicle regions on different frames of follicle images are regarded as the same follicle marking regions.

[0025] Furthermore, after obtaining the marked area of ​​the same follicle, the following steps are also included:

[0026] According to the morphological characteristics of each follicle-marked area on the last frame of the follicle image and the possibility of corresponding to the same follicle, the quality of each follicle-marked area is obtained, and the optimal follicle area is screened out.

[0027] Furthermore, the method for obtaining the degree of excellence includes:

[0028] For the last frame of follicle image, the number of pixels and the probability of the same follicle in each marked area of ​​the same follicle are fused to obtain the excellence of the corresponding marked area of ​​the same follicle. The number of pixels and the probability of the same follicle are positively correlated with the excellence.

[0029] Furthermore, the method for obtaining the optimal follicle region includes:

[0030] The area with the largest value among all the areas marked with the same follicle is selected, and the area marked with the same follicle is taken as the optimal follicle area.

[0031] The present invention also proposes a reproductive medicine nursing image assisted identification system, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements any one of the steps of the reproductive medicine nursing image assisted identification method.

[0032] The present invention has the following beneficial effects:

[0033] The present invention takes into account that the interior of the follicle region is an echo-free area, which has a significant difference in grayscale value from the surrounding low or high echo areas, and both present a quasi-circular follicle shape. For any frame of follicle image, the connected domain of the follicle image is obtained according to the grayscale distribution of pixels on the follicle image; since the circular boundary is continuous and smooth, the possibility of the suspected follicle region of each connected domain is obtained according to the edge morphological characteristics of each connected domain; according to the possibility of the suspected follicle region, the suspected follicle region of each frame of the follicle image is screened out, the area most likely to be the follicle is determined, and the scope of subsequent processing is narrowed; considering that the changes of the same follicle are usually continuous and Regularly, according to the position distribution and morphological characteristics of suspected follicle areas between follicle images of different frames, multiple groups of matching suspected follicle areas are obtained; according to the position change characteristics and morphological change characteristics of each group of matching suspected follicle areas, the possibility of the same follicle in each group of matching suspected follicle areas is obtained, and the change similarity of each group of matching suspected follicle areas is quantified; according to the possibility of the same follicle, the same follicle marked area on the follicle images of different frames is screened out, which is helpful to understand the connection between the same follicle in different frames; the present invention improves the analysis accuracy of follicle development by accurately identifying the same follicle area corresponding to different frames of follicle images. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0035] Figure 1 A flowchart of a reproductive medicine nursing image-assisted recognition method provided by an embodiment of the present invention;

[0036] Figure 2 A schematic diagram of an ovarian follicle image provided by an embodiment of the present invention;

[0037] Figure 3 A flow chart of a method for obtaining a matching suspected follicle region provided by one embodiment of the present invention;

[0038] Figure 4 A flow chart of a method for obtaining the possibility of a same follicle provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0039] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the specific implementation method, structure, features and effects of a reproductive medicine nursing image-assisted recognition method and system proposed by the present invention in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0040] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0041] The specific scheme of the reproductive medicine nursing image-assisted recognition method and system provided by the present invention is described in detail below with reference to the accompanying drawings.

[0042] See also Figure 1 , which shows a method flow chart of a reproductive medicine nursing image-assisted recognition method provided by an embodiment of the present invention, specifically comprising:

[0043] Step S1: Acquire multiple frames of follicle images of a patient preparing for pregnancy in time sequence.

[0044] In an embodiment of the present invention, in order to use medical assistance to help infertile couples conceive, it is necessary to monitor the growth of follicles in patients who are preparing for pregnancy. First, according to the menstrual cycle stage of a patient and her clinical manifestations, the frequency of transvaginal ultrasound examination of the patient is planned. Usually, after the 8th day of the menstrual cycle of the patient who is preparing for pregnancy, the B-ultrasound probe is placed in the vagina of the patient who is preparing for pregnancy to observe the entire pelvic structure. In order to reduce subsequent errors, the probe position is ensured to be the same each time, and several transvaginal ultrasound examinations at the same position are obtained to analyze the follicle status and development degree; multiple frames of follicle images of the patient who is preparing for pregnancy are obtained in time sequence.

[0045] It should be noted that, in the embodiment of the present invention, the number of frames of follicle images acquired is 3 or more, and analysis is continuously performed until the follicles mature, wherein the same method is used for the analysis of subsequent acquisitions of multiple frames of follicle images.

[0046] It should be noted that, in one embodiment of the present invention, the timing interval is 1 day, that is, transvaginal ultrasound monitoring is performed at the same time every day; in other embodiments of the present invention, the size of the timing interval can be set according to the specific circumstances, and is not limited or elaborated here.

[0047] Step S2: For any frame of follicle image, the connected domain of the follicle image is obtained according to the grayscale distribution of pixels on the follicle image; the possibility of the suspected follicle area of ​​each connected domain is obtained according to the edge morphological characteristics of each connected domain; according to the possibility of the suspected follicle area, the suspected follicle area of ​​each frame of follicle image is screened out.

[0048] Since the follicle area is an echo-free area, the gray value is significantly different from the surrounding low or high echo areas, and both appear in a round bubble shape, such as Figure 2 , shows a schematic diagram of a follicle image, and the part marked with a digital serial number is displayed as a follicle; therefore, for any frame of the follicle image, the connected domain of the follicle image is obtained according to the grayscale distribution of the pixels on the follicle image, which reduces the interference of noise on the analysis and helps to perform separate processing and analysis of each area later.

[0049] It should be noted that, in one embodiment of the present invention, the method for obtaining the connected domain includes: performing edge detection on each frame of the follicle image, obtaining the edge curve in each frame of the follicle image, and taking the area surrounded by the closed edge curve as the connected domain. It should be noted that the specific edge detection is a technical means well known to those skilled in the art and will not be described in detail here.

[0050] Since the follicle region is quasi-circular in shape, the circular boundary is continuous and smooth, and the direction changes between adjacent edge pixels are small and uniform, the possibility of the suspected follicle region of each connected domain is obtained according to the edge morphological characteristics of each connected domain.

[0051] Preferably, in one embodiment of the present invention, the method for obtaining the possibility of a suspected follicle region includes:

[0052] The 8-chain code of the edge pixel point can represent the boundary direction of the connected domain. For each connected domain, the difference between the 8-chain code value of each edge pixel point and the previous adjacent edge pixel point is obtained as the chain code difference;

[0053] According to the absolute value of the average chain code difference between all edge pixels and the previous adjacent edge pixels, as well as the fluctuation characteristics of the chain code difference, the possibility of the suspected follicle area of ​​each connected domain is obtained. The absolute value of the average chain code difference and the fluctuation characteristics of the chain code difference are negatively correlated with the possibility of the suspected follicle area.

[0054] It should be noted that, in one embodiment of the present invention, the overall change level of the edge of the connected domain is represented by calculating the mean of the chain code differences between all edge pixels and the previous adjacent edge pixels, that is, the average chain code difference. The smaller the absolute value of the average chain code difference, the closer the chain code values ​​between adjacent edge pixels, the smaller the directional change between adjacent edge pixels, the smoother the edge, the more likely it is a follicle area, and the greater the possibility of a suspected follicle area, showing a negative correlation.

[0055] In one embodiment of the present invention, the fluctuation characteristics are represented by calculating the variance. The larger the variance, the larger the fluctuation characteristics. The smaller the fluctuation characteristics, the more uniform the change in the chain code difference, the more likely it is a follicle area, and the greater the possibility of a suspected follicle, showing a negative correlation. In other embodiments of the present invention, methods such as range and standard deviation can also be used to represent the fluctuation characteristics. The specific means are technical means well known to those skilled in the art and will not be elaborated here.

[0056] In one embodiment of the present invention, taking a connected domain as an example, the formula for the possibility of a suspected follicle region is expressed as:

[0057] ;

[0058] in, The suspected follicle region possibility is indicated by the connected domain; Represents the average chain code difference between all edge pixels on the connected domain and the previous adjacent edge pixels; Represents the absolute value of the average chain code difference between all edge pixels on the connected domain and the previous adjacent edge pixels; Represents the variance of the chain code difference between all edge pixels on the connected domain and the previous adjacent edge pixels, that is, the fluctuation feature; Represents an exponential function with a natural constant as its base.

[0059] In the formula for the likelihood of a suspected follicle region, For negative correlation mapping, the smaller the absolute value of the average chain code difference, the smaller the fluctuation feature, the more uniform the edge area changes between adjacent edge pixels, the smoother the edge changes of the connected domain, the more likely it is the follicle area, and the greater the possibility of suspected follicle area; on the contrary, the larger the absolute value of the average chain code difference, the larger the fluctuation feature, the more unstable the edge area changes between adjacent edge pixels, the more likely there is bending or turning in the edge changes of the connected domain, and the smaller the possibility of suspected follicle area.

[0060] It should be noted that in other embodiments of the present invention, the possibility of a suspected follicle area in a connected domain can also be reflected by analyzing the circularity of the connected domain. The larger the circularity, the more likely it is a suspected follicle area, and the greater the possibility of a suspected follicle area. The specific means are technical means well known to those skilled in the art and will not be elaborated here.

[0061] By analyzing the possibility of suspected follicle regions in each connected domain, the connected domain with higher possibility is more likely to be considered as a follicle region, which can effectively identify and locate follicles and help improve the accuracy and reliability of follicle regions. Therefore, according to the possibility of suspected follicle regions, the suspected follicle regions of each frame of follicle image are screened out.

[0062] Preferably, in one embodiment of the present invention, the method for obtaining the suspected follicle region includes:

[0063] For any frame of follicle image, if the probability of the suspected follicle region in the connected domain is greater than or equal to the preset probability threshold, the corresponding connected domain is taken as the suspected follicle region.

[0064] It should be noted that, in one embodiment of the present invention, the size of the preset possibility threshold is 0.7; in other embodiments of the present invention, the size of the preset possibility threshold can be set according to specific circumstances, and is not limited or elaborated here.

[0065] Step S3: obtaining multiple groups of matching suspected follicle regions according to the position distribution and morphological characteristics of the suspected follicle regions between different frames of follicle images; obtaining the possibility that each group of matching suspected follicle regions is the same follicle according to the position change characteristics and morphological change characteristics of each group of matching suspected follicle regions.

[0066] The development of follicles is a continuous process. From primordial follicles to preovulatory follicles, each stage of development has its own specific morphological characteristics and position distribution. However, the displacement and morphological changes of follicles in a short period of time are relatively stable. Therefore, by observing the position and morphological characteristics of the suspected follicle area in different frames of follicle images, the corresponding follicle area can be matched according to its continuity and change rules. According to the position distribution and morphological characteristics of the suspected follicle area between different frames of follicle images, multiple groups of matching suspected follicle areas are obtained.

[0067] Preferably, in one embodiment of the present invention, the method for obtaining the matching suspected follicle region can be found in Figure 3 , which shows a flow chart of a method for obtaining a matching suspected follicle region, including:

[0068] Step S301: Obtain the centroid coordinates of each suspected follicle region in each frame of follicle image.

[0069] In order to ensure the consistency and comparability of the analysis between different frames of follicle images, the left edge of each frame of follicle image is used as the vertical axis of the coordinate system, and the lower edge perpendicular to the vertical axis is used as the horizontal axis of the coordinate system. The same operation is used to construct the coordinate system for the follicle image. In one embodiment of the present invention, the centroid coordinates are obtained by calculating the mean of the position coordinates of all pixels in the suspected follicle area as the centroid coordinates.

[0070] Step S302: In each frame of the follicle image, a neighborhood range is constructed with the centroid coordinates of each suspected follicle area in the previous frame of the follicle image as the center. If the number of pixel points in the suspected follicle area in the corresponding neighborhood range is greater than the number of pixel points in the suspected follicle area corresponding to the previous frame of the follicle image, the corresponding suspected follicle area in the neighborhood range is used as a candidate suspected follicle area; the candidate suspected follicle area with the smallest relative distance between the corresponding centroid coordinates and the center position of the neighborhood range is selected as the same type of suspected follicle area as the suspected follicle area corresponding to the previous frame of the follicle image.

[0071] The follicles in different frames of follicle images are constantly developing. During the short-term development of the follicles, the changes in displacement and morphology are relatively stable and the displacement rate is relatively small. Therefore, the number of pixels must be greater than the number of pixels in the previous frame of the follicle image corresponding to the suspected follicle area, and the relative distance is relatively small, the more likely it is that the follicles are developing and changing in the same type.

[0072] It should be noted that, in one embodiment of the present invention, the method for obtaining the relative distance is to calculate the Euclidean distance; in other embodiments of the present invention, the relative distance can also be obtained by using existing distance calculation methods such as Manhattan distance. The specific means are technical means well known to those skilled in the art and will not be elaborated here.

[0073] It should be noted that, in one embodiment of the present invention, the neighborhood range is obtained by taking the centroid coordinates as the center and constructing a radius of The circular range is used as the neighborhood range, wherein the radius can be set by the implementer according to the specific situation. If there is no suspected follicle area within the circular range, it means that there is no match. In other embodiments of the present invention, the size of the neighborhood range can be set according to the specific situation, which is not limited or elaborated here.

[0074] Step S303: If each suspected follicle region has similar suspected follicle regions in different frames of follicle images, a group of matching suspected follicle regions is formed.

[0075] Due to the influence of operation and environment, as well as the interference of B-ultrasound itself, there may be similar round follicle-shaped interference items in the process of follicular development. Considering that the interference items appear randomly and there will be no continuous change trend, each suspected follicle area is selected to have similar suspected follicle areas in different frames of follicle images for matching analysis.

[0076] The same follicle may have a certain displacement and growth pattern in different frame images. The follicle will gradually move toward the ovarian wall, and the displacement distance will not be large in a short time. The follicle will gradually grow larger over time. The growth rate and displacement rate of the follicle are relatively stable in a short time. The follicle changes continuously over time. Matching the position change characteristics and morphological change characteristics between suspected follicle areas helps to reflect the changing trend of the follicle. Therefore, the possibility of the same follicle in each group of matching suspected follicle areas is obtained based on the position change characteristics and morphological change characteristics of each group of matching suspected follicle areas.

[0077] Preferably, in one embodiment of the present invention, the method for obtaining the probability of the same follicle can be found in Figure 4 , which shows a flow chart of a method for obtaining the possibility of a same follicle, including:

[0078] Step S401: For any group of matched suspected follicle regions, the ratio of the number of pixels in the matched suspected follicle region between each frame of follicle image and its previous frame of follicle image is obtained as the morphological increase rate of the matched suspected follicle region corresponding to each frame of follicle image.

[0079] The number of pixels in the matching suspected follicle area can reflect the size of the morphological area of ​​the matching suspected follicle area. The more pixels there are, the larger the area is, and the larger the morphology is. The ratio of the number of pixels in the matching suspected follicle area between each frame of the follicle image and the previous frame of the follicle image can reflect the morphological changes of the matching suspected follicle area. The larger the ratio, the smaller the number of pixels in the matching suspected follicle area of ​​the previous frame, and the greater the morphological increase rate.

[0080] In one embodiment of the present invention, the formula of the morphology increase rate can be expressed as: ,in, Indicates The frame follicle image corresponds to the morphological enlargement rate of the matching suspected follicle area; Indicates The frame follicle image corresponds to the number of pixels in the suspected follicle area; Indicates The frame follicle image corresponds to the number of pixels in the suspected follicle area; Indicates the frame number of the follicle image.

[0081] Step S402: obtaining the difference of the centroid coordinates in the matching suspected follicle region between each frame of follicle image and the previous frame of follicle image as the displacement vector of the matching suspected follicle region corresponding to each frame of follicle image.

[0082] By calculating the difference in centroid coordinates, the displacement of the follicle in the image can be accurately quantified, and the displacement vector obtained reflects the growth speed and direction of the follicle. The larger the difference in centroid coordinates, the greater the distance between the centroid coordinates, and the faster the follicle grows.

[0083] It should be noted that, in one embodiment of the present invention, the formula of the displacement vector is expressed as: ;in, Indicates The frame follicle image corresponds to the displacement vector of the matching suspected follicle area; Indicates The frame follicle image corresponds to the centroid abscissa of the matching suspected follicle area; Indicates The frame follicle image corresponds to the centroid abscissa of the matching suspected follicle area; Indicates The frame follicle image corresponds to the centroid ordinate of the matching suspected follicle area; Indicates The frame follicle image corresponds to the centroid ordinate of the matching suspected follicle area.

[0084] Step S403: According to the difference in morphological increase rate and displacement vector of the corresponding matching suspected follicle area between adjacent frame follicle images, the possibility of the same follicle in the corresponding group matching suspected follicle area is obtained. The difference in morphological increase rate and displacement vector are negatively correlated with the possibility of the same follicle.

[0085] The difference in morphological enlargement rate can reflect the morphological changes of the matching suspected follicle areas between adjacent frame follicle images. The greater the difference in morphological enlargement rate, the more dissimilar the morphological changes, the worse the stability, and the smaller the possibility of the same follicle, showing a negative correlation.

[0086] It should be noted that, since the displacement vector is a vector, including size and direction, in order to analyze the displacement changes of the suspected follicle area matched between adjacent frames of follicle images, in one embodiment of the present invention, the difference in the modulus length of the displacement vector that matches the suspected follicle area between adjacent frames of follicle images and the angle of the displacement vector are analyzed to represent the difference in the displacement vector. The greater the difference in the modulus length of the displacement vector and the larger the angle of the displacement vector, the more dissimilar the displacement changes of the suspected follicle area matched between adjacent frames of follicle images and the smaller the possibility of the same follicle, that is, the difference in the displacement vector is negatively correlated with the possibility of the same follicle.

[0087] In one embodiment of the present invention, for any set of matching suspected follicle regions, the formula for the possibility of the same follicle is expressed as:

[0088] ;

[0089] in, Indicates the probability of a group of matching suspected follicle regions being the same follicle; Indicates The frame follicle image corresponds to the morphological enlargement rate of the matching suspected follicle area; Indicates The frame follicle image corresponds to the morphological enlargement rate of the matching suspected follicle area; Indicates The frame follicle image corresponds to the displacement vector of the matching suspected follicle area; Indicates The frame follicle image corresponds to the displacement vector of the matching suspected follicle area; Indicates Frame follicle image and The angle between the displacement vectors of the corresponding matching suspected follicle regions between the frame follicle images; Indicates Frame follicle image and The cosine value of the angle between the displacement vectors of the suspected follicle regions corresponding to the frame follicle images; Indicates The modulus of the displacement vector of the frame follicle image corresponding to the matched suspected follicle area; Indicates The modulus of the displacement vector of the frame follicle image corresponding to the matched suspected follicle area; Indicates the frame number of the follicle image.

[0090] In the formula for the probability of the same follicle, It means that the overall changes between the follicle images of all adjacent frames are analyzed by averaging, and the angle is mapped by the cosine value function. The larger the angle, the smaller the cosine value of the angle, and the smaller the possibility of the same follicle; Indicates the calculation Frame follicle image and The absolute value of the difference in the modulus length of the displacement vector of the suspected follicle area corresponding to the frame follicle image is as follows: the smaller the absolute value of the difference in modulus length is, the closer the displacement vector moves, and the greater the possibility of the same follicle; Indicates the calculation Frame follicle image and The absolute values ​​of the difference in the morphological increase rates of the corresponding matching suspected follicle areas between the frame follicle images are as follows. The larger the absolute value of the difference, the more inconsistent the morphological increase rates are, and the smaller the possibility of the same follicle is.

[0091] Step S4: According to the possibility of the same follicle, the same follicle marked area on the follicle images of different frames is screened out.

[0092] By analyzing the possibility of the same follicle in each group of matching suspected follicle areas, the connected domain with higher possibility is more likely to be considered as the area of ​​the same follicle change, which can effectively identify and locate the follicle, and help improve the accuracy and reliability of the follicle area. Therefore, according to the possibility of the same follicle, the same follicle marked area on different frames of follicle images is screened out.

[0093] Preferably, in one embodiment of the present invention, the method for obtaining the same follicle marking area includes:

[0094] If the probability of a group of matching suspected follicle regions being the same follicle is greater than or equal to a preset same follicle threshold, the corresponding matching suspected follicle regions on different frames of follicle images are regarded as the same follicle marking regions.

[0095] It should be noted that, in one embodiment of the present invention, the size of the preset follicle threshold is 0.75; in other embodiments of the present invention, the size of the preset follicle threshold is a technical means well known to those skilled in the art, which will not be limited or elaborated herein.

[0096] It should be noted that after obtaining the same follicle marked area, it also includes: according to the morphological characteristics of each same follicle marked area on the last frame of the follicle image and the possibility of corresponding to the same follicle, the quality of each same follicle marked area is obtained, and the optimal follicle area is screened out.

[0097] Excellent follicles usually have a round and full appearance and develop relatively stably. Therefore, the larger the morphological characteristics and the more consistent the changes between different frames, the greater the possibility of the same follicle and the greater the degree of excellence. Therefore, based on the morphological characteristics of each follicle marked area on the last frame of the follicle image and the corresponding possibility of the same follicle, the excellence of each follicle is obtained and the optimal follicle area is screened out.

[0098] Preferably, in one embodiment of the present invention, the method for obtaining the degree of excellence includes:

[0099] For the last frame of follicle image, the number of pixels and the probability of the same follicle in each marked area of ​​the same follicle are fused to obtain the excellence of the corresponding marked area of ​​the same follicle. The number of pixels and the probability of the same follicle are positively correlated with the excellence.

[0100] It should be noted that, in some embodiments of the present invention, the number of pixels in the follicle area and the possibility of the same follicle can be fused by multiplication or addition. The larger the number of pixels, the greater the possibility of the same follicle and the greater the quality of the follicle. The specific means are technical means well known to those skilled in the art and will not be elaborated here.

[0101] In one embodiment of the present invention, for the last frame of follicle image, the formula for the degree of excellence is expressed as:

[0102] ;

[0103] in, Indicates The quality of the labeled area of ​​the same follicle; Represents the last frame of follicle image No. The number of pixels in the same follicle marking area; Indicates The probability of the same follicle area in the same follicle marking area.

[0104] In the formula for excellence, the greater the number of pixels, the greater the possibility of the same follicle area, the more stable the follicle development, and the greater the excellence.

[0105] Preferably, in one embodiment of the present invention, the method for obtaining the optimal follicle region includes:

[0106] The area with the largest value among all the areas marked with the same follicle is selected, and the area marked with the same follicle is taken as the optimal follicle area.

[0107] It should be noted that, in another embodiment of the present invention, the same follicle marking area and the optimal follicle area can more intuitively reflect the growth and development status of the follicle. After obtaining the same follicle marking area and the optimal follicle area on different frames of follicle images, the follicles can be identified and monitored more accurately, including: performing local histogram homogenization processing on the same follicle marking area on each frame of the follicle image to obtain an enhanced follicle image, so that the outline and internal structure of the follicle area are more prominent in the image, while improving its brightness and enhancing its information display intensity, and further combining with the position of the optimal follicle area, the growth and development of the follicles between different frames of images can be more clearly understood, thereby improving the recognition and monitoring effects of the follicles.

[0108] In summary, the present invention obtains the connected domain of the follicle image; obtains the possibility of the suspected follicle area of ​​each connected domain according to the edge morphological characteristics of each connected domain, and screens out the suspected follicle area of ​​each frame of the follicle image; analyzes the position distribution and morphological characteristics of the suspected follicle area between different frames of follicle images, obtains multiple groups of matching suspected follicle areas, and obtains the possibility of the same follicle in each group of matching suspected follicle areas; screens out the same follicle marked area on different frames of follicle images; the present invention improves the analysis accuracy of follicle development by accurately identifying the same follicle area corresponding to different frames of follicle images.

[0109] The present invention also proposes a reproductive medicine nursing image assisted identification system, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements any one of the steps of a reproductive medicine nursing image assisted identification method.

[0110] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0111] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A reproductive medicine nursing image-assisted recognition method, characterized in that: The method comprises: Acquire multiple frames of follicle images of patients preparing for pregnancy in time sequence; For any frame of follicle image, the connected domain of the follicle image is obtained according to the grayscale distribution of the pixels on the follicle image; the possibility of the suspected follicle region of each connected domain is obtained according to the edge morphological characteristics of each connected domain; and the suspected follicle region of each frame of follicle image is screened out according to the possibility of the suspected follicle region; According to the position distribution and morphological characteristics of the suspected follicle regions between different frames of follicle images, multiple groups of matching suspected follicle regions are obtained; according to the position change characteristics and morphological change characteristics of each group of matching suspected follicle regions, the possibility of the same follicle in each group of matching suspected follicle regions is obtained; According to the same follicle possibility, the same follicle marked area on the follicle images of different frames is screened out; The method for obtaining the matching suspected follicle region comprises: Obtain the centroid coordinates of each suspected follicle region in each frame of follicle image; In each frame of follicle image, a neighborhood range is constructed with the centroid coordinates of each suspected follicle region of the previous frame of follicle image as the center. If the number of pixel points in the suspected follicle region in the corresponding neighborhood range is greater than the number of pixel points in the suspected follicle region corresponding to the previous frame of follicle image, the corresponding suspected follicle region in the neighborhood range is used as a candidate suspected follicle region; the candidate suspected follicle region with the smallest relative distance between the corresponding centroid coordinates and the center position of the neighborhood range is selected as the same type of suspected follicle region corresponding to the suspected follicle region of the previous frame of follicle image; If each suspected follicle region has similar suspected follicle regions in different frames of follicle images, a group of matching suspected follicle regions is formed; The method for obtaining the probability of the same follicle comprises: For any group of matched suspected follicle regions, the ratio of the number of pixels in the matched suspected follicle region between each frame of follicle image and the previous frame of follicle image is obtained as the morphological enlargement rate of the matched suspected follicle region corresponding to each frame of follicle image; Obtaining the difference in centroid coordinates in the matching suspected follicle region between each frame of follicle image and the previous frame of follicle image as a displacement vector of the matching suspected follicle region corresponding to each frame of follicle image; According to the difference in morphological enlargement rate and displacement vector of the corresponding matching suspected follicle area between adjacent frames of follicle images, the possibility of the same follicle in the corresponding group of matching suspected follicle areas is obtained. The difference in morphological enlargement rate and displacement vector are negatively correlated with the possibility of the same follicle.

2. The reproductive medicine nursing image-assisted recognition method according to claim 1, characterized in that: The method for obtaining the possibility of the suspected follicle area includes: For each connected domain, obtain the difference between the 8-chain code value of each edge pixel and the previous adjacent edge pixel as the chain code difference; According to the absolute value of the average chain code difference between all edge pixels and the previous adjacent edge pixels, as well as the fluctuation characteristics of the chain code difference, the possibility of the suspected follicle area of ​​each connected domain is obtained. The absolute value of the average chain code difference and the fluctuation characteristics of the chain code difference are negatively correlated with the possibility of the suspected follicle area.

3. The reproductive medicine nursing image-assisted recognition method according to claim 1, characterized in that: The method for obtaining the suspected follicle region comprises: For any frame of follicle image, if the probability of the suspected follicle region in the connected domain is greater than or equal to the preset probability threshold, the corresponding connected domain is taken as the suspected follicle region.

4. The reproductive medicine nursing image-assisted recognition method according to claim 1, characterized in that: The method for obtaining the same follicle marked area comprises: If the probability of a group of matching suspected follicle regions being the same follicle is greater than or equal to a preset same follicle threshold, the corresponding matching suspected follicle regions on different frames of follicle images are regarded as the same follicle marking regions.

5. The reproductive medicine nursing image-assisted recognition method according to claim 1, characterized in that: After obtaining the same follicle marked area, it also includes: According to the morphological characteristics of each follicle-marked area on the last frame of the follicle image and the possibility of corresponding to the same follicle, the quality of each follicle-marked area is obtained, and the optimal follicle area is screened out.

6. A reproductive medicine nursing image-assisted recognition method according to claim 5, characterized in that: The method for obtaining the degree of excellence includes: For the last frame of follicle image, the number of pixels and the probability of the same follicle in each marked area of ​​the same follicle are fused to obtain the excellence of the corresponding marked area of ​​the same follicle. The number of pixels and the probability of the same follicle are positively correlated with the excellence.

7. The reproductive medicine nursing image-assisted recognition method according to claim 5, characterized in that: The method for obtaining the optimal follicle region comprises: The area with the largest value among all the areas marked with the same follicle is selected, and the area marked with the same follicle is taken as the optimal follicle area.

8. A reproductive medicine nursing image-assisted recognition system, the system comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the reproductive medicine nursing image-assisted recognition method as described in any one of claims 1 to 7 are implemented.

Images