Blood smear evaluation method, device and system based on medical examination

Abstract

The invention relates to the technical field of medical image processing, in particular to a blood smear evaluation method, device and system based on medical examination. According to the method, blood cell images of all areas of a blood smear are obtained in a traversal scanning mode, and an image set is constructed by taking a column as a unit. The cell aggregation degree of each blood cell image can be accurately and effectively quantified by screening out the red blood cell central area and determining the first blood cell area and the second blood cell area for joint analysis. And constructing a cell aggregation degree sequence and analyzing the uniformity in the cell aggregation degree sequence so as to obtain the observation difficulty of each image set, and screening out an observable area, so that accurate blood smear quality evaluation is realized. According to the method, the cell aggregation degree in the image is effectively quantified in a traversal scanning image acquisition mode, and a correct observable area is divided according to the special form of the blood smear, so that accurate blood smear quality evaluation is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of medical image processing, in particular to a blood smear evaluation method, device and system based on medical examination. BACKGROUND

[0002] Microscopic examination is the basic method of blood cell morphology examination, mainly used for red blood cell, white blood cell and platelet morphology examination, and can also be used for white blood cell and platelet quantity evaluation, and detection of possible parasites in blood. Through cell morphology change, abnormal cells can be found to determine some blood diseases. Blood smears are artificially made, and their quality is uneven, which has a significant impact on subsequent detection.

[0003] Blood smears are usually made by using a glass push piece to quickly and smoothly push a drop of blood to the other end of the slide, and the thickness of the blood film formed varies. The thickness of the blood film on one side is high, and the thickness of the blood film on the other side is low. In the area where the blood film is thick, the degree of cell aggregation is high, and in the area where the blood film is thin, the number of cells is insufficient. The quality of blood smears should be judged in the area where blood cells are easy to observe, rather than evaluating the uniformity of the whole smear. The existing technology mainly screens the observable area of the blood smear by comparing the standard under the microscope, which is easy to cause misjudgment and omission, thereby affecting the blood smear quality evaluation result. SUMMARY

[0004] In order to solve the technical problem that the positioning of the observable area in the blood smear is not accurate in the prior art, thereby affecting the blood smear quality evaluation result, the purpose of the present application is to provide a blood smear evaluation method, device and system based on medical examination, and the technical solution adopted is as follows: The present application provides a blood smear evaluation method based on medical examination, which comprises: Image scanning is performed on the blood smear vertically from the tail to the head, and the blood cell images scanned vertically each time form an image set in order; all image sets of the blood area on the blood smear are obtained; The threshold segmentation algorithm is used to obtain the suspected blood cell nucleus area and the suspected cytoplasm area in the blood cell image; the red blood cell center area is screened according to the size and edge curvature of the suspected blood cell nucleus area; the red blood cell center area and the adjacent suspected cytoplasm area are combined as a first blood cell area, and the other suspected cytoplasm area is a second blood cell area; the cell aggregation degree of each blood cell image is obtained according to the edge length and gray value of the first blood cell area and the second blood cell area; The cell aggregation degree in each image set forms a cell aggregation degree sequence according to the position sequence of the blood cell images; a cell aggregation degree uniformity of the cell aggregation degree sequence is obtained; an observation difficulty of each image set is obtained according to the cell aggregation degree uniformity and the cell aggregation degree; and an observable region is screened out in the scanning region according to the observation difficulty. The blood smear quality is obtained according to the area of the observable region and the shape integrity of the blood cell region in the observable region.

[0005] Further, the obtaining of the suspected blood cell nucleus region and the suspected cytoplasm region in the blood cell image comprises: The blood cell image is processed by using a threshold segmentation algorithm to obtain a binary image; in the binary image, a connected domain composed of a type of pixel points with the highest gray scale is a suspected nucleus region, and a connected domain composed of another type of pixel points is a suspected cytoplasm region.

[0006] Further, the screening method of the red blood cell center region comprises: The suspected nucleus region smaller than the preset area threshold is taken as a second suspected nucleus region, the curvature of the second suspected nucleus region is obtained and compared with the curvature of a standard circle, and the second suspected nucleus region with a curvature error within a preset error range from the curvature of the standard circle is taken as the red blood cell center region.

[0007] Further, the obtaining method of the cell aggregation degree comprises: The first blood cell region and the second blood cell region are taken as analysis regions, the gray mean value of the analysis regions in the blood cell image is obtained, the gray mean value is negatively correlated and mapped, and then multiplied by the edge length of the analysis region to obtain the initial aggregation degree of each analysis region, and the initial aggregation degrees of all the analysis regions in the blood cell image are accumulated and taken as the cell aggregation degree.

[0008] Further, the obtaining method of the cell aggregation degree uniformity comprises: According to the element value of the central position of the cell aggregation degree sequence, a comparison element value is obtained, the cell aggregation degree difference between the comparison element value and each element in the cell aggregation degree sequence is obtained, the cell aggregation degree differences of all the elements are negatively correlated and mapped, and then accumulated to obtain the cell aggregation degree uniformity.

[0009] Further, the obtaining method of the observation difficulty comprises: The product of the cell aggregation degree uniformity and the average cell aggregation degree in the image set is normalized to obtain the observation difficulty.

[0010] Further, the obtaining method of the shape integrity of the blood cell region in the observable region comprises: The first blood cell region and the second blood cell region in the observable region are taken as to-be-analyzed regions, convex hull detection is performed on each to-be-analyzed region, a convex hull edge of each to-be-analyzed region is obtained, an edge point in the convex hull edge which is not an edge of the to-be-analyzed region is taken as a difference edge point, a ratio between a number of the difference edge points and a number of the edge of the to-be-analyzed region is taken as an incompleteness, the incompleteness is negatively correlated mapping, an initial shape completeness of each to-be-analyzed region is obtained, a to-be-analyzed region with an initial shape completeness greater than a preset first completeness threshold is taken as a complete blood cell region, and a proportion of the number of the complete blood cell region in all to-be-analyzed regions is taken as the shape completeness of the blood cell region in the observable region.

[0011] Further, the blood smear quality acquisition method comprises: If the area proportion is less than a preset area proportion threshold, it is determined that the blood smear quality is substandard. If the area proportion is not less than the preset area proportion threshold, a shape completeness of the blood cell region in the observable region is obtained, if the shape completeness is greater than a preset second completeness threshold, it is determined that the blood smear quality is standard, and if the shape completeness is not greater than the preset second completeness threshold, it is determined that the blood smear quality is substandard.

[0012] The application further provides a blood smear evaluation device based on medical examination, which comprises: The image acquisition module is configured to perform image scanning on the blood smear in a direction perpendicular to the tail-to-head direction, and the blood cell images scanned in each time of the perpendicular scanning are sequentially arranged to form an image set. The cell aggregation degree acquisition module is configured to obtain a suspected blood cell nucleus region and a suspected cytoplasm region in the blood cell image by using a threshold segmentation algorithm, screen a red blood cell center region according to the size and edge curvature of the suspected blood cell nucleus region, and combine the red blood cell center region and the adjacent suspected cytoplasm region as a blood cell region. The observable region positioning module is configured to arrange the cell aggregation degrees in each image set in a sequence according to the position sequence of the blood cell image, obtain the cell aggregation degree uniformity of the cell aggregation degree sequence, and obtain the observation difficulty of each image set according to the cell aggregation degree uniformity and the cell aggregation degree. The blood smear quality evaluation module is configured to obtain the blood smear quality according to the area of the observable region and the shape completeness of the blood cell region in the observable region.

[0013] The application further provides a blood smear evaluation system based on medical examination, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the steps of any one of the blood smear evaluation methods based on medical examination.

[0014] The application has the following advantages: The embodiment of the application obtains blood cell images of all areas of the blood smear in a traversal scanning manner, and constructs an image set in column units. The application considers that the observable area should have the characteristics of low blood cell aggregation and easy observation, and therefore quantifies the cell aggregation degree in each blood cell image. Considering that red blood cells in the blood have the characteristics of concave in the middle and convex around, which causes the middle gray value to be obviously smaller than the surrounding gray value, directly using a threshold segmentation algorithm to determine the cell region will cause the determined red blood cell region to lack information of the middle position. Therefore, the application jointly analyzes the first blood cell region and the second blood cell region by screening out the central region of the red blood cell, and can accurately and effectively quantify the cell aggregation degree of each blood cell image. Because the blood smear has a specific smearing manner during smearing, and further forms a special blood film thickness distribution, the application further constructs a cell aggregation degree sequence and analyzes the uniformity thereof, and further obtains the observation difficulty of each image set, and can screen out the observable area, i.e., the area with appropriate blood aggregation degree and uniformity. Thus, accurate blood smear quality evaluation is realized. BRIEF DESCRIPTION OF DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the embodiments of the application or the prior art, and the advantages thereof, the following will briefly introduce the drawings needed in the embodiments or the prior art description. Obviously, the drawings in the following description are only some embodiments of the application, and for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 A blood smear evaluation method based on medical examination provided by an embodiment of the application is shown in the flowchart; Figure 2 A blood cell image gray scale diagram provided by an embodiment of the application is shown in the diagram; Figure 3 A binary image provided by an embodiment of the application is shown in the diagram; Figure 2 Figure 4 A diagram showing the blood diffusion direction in the blood smearing process provided by an embodiment of the application is shown in the diagram; Figure 5 ​A schematic diagram of an observable area of a blood smear provided by one embodiment of the present application. DETAILED DESCRIPTION

[0017] In order to further clarify the technical means and effects taken by the present application to achieve the predetermined inventive objectives, the following describes in detail the specific implementation, structure, features and effects of a blood smear evaluation method, device and system based on medical examination according to the present application, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0018] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs.

[0019] The following specifically describes the specific scheme of a blood smear evaluation method, device and system based on medical examination provided by the present application in combination with the accompanying drawings.

[0020] Please refer to Figure 1 which shows a flowchart of a blood smear evaluation method based on medical examination provided by one embodiment of the present application, the method comprising: Step S1: image scanning is performed on the blood smear in a direction perpendicular to the tail to the head, and the blood cell images scanned vertically each time form an image set in order; all image sets of the blood region on the blood smear are obtained.

[0021] In the embodiment of the present application, the blood smear is the blood smear after venous blood sampling and processing by an automatic smear machine, which is placed under a microscope to collect images. A camera specially used for a microscope is used to capture the eyepiece image of the microscope to form a real-time image after being installed in the eyepiece. In the embodiment of the present application, image scanning is performed from the tail of the blood smear to the head along the vertical direction of the smear, and each time the scanning ends at the edge of the blood film to form an image set of blood cell images. The blood cell images in the set are arranged in the order of the position in the smear according to the order of image collection. After each scanning ends, the blood smear is moved 3mm towards the head to continue the vertical collection, and then all image sets of the blood region on the blood smear are obtained through S-shaped vertical scanning. Each image set can be regarded as the information of a column of regions in the blood region.

[0022] It should be noted that when the blood smear is scanned, the other regions without blood regions do not need to be scanned, and the method for obtaining the blood region comprises: obtaining the overall image of the blood smear, performing Otsu threshold segmentation under the R channel of the overall image, and the obtained region is the blood region.

[0023] Step S2: facilitate the threshold segmentation algorithm to obtain the suspected blood cell nucleus region and suspected cytoplasm region in the blood cell image; according to the size of the suspected nucleus region and the edge curvature, the red blood cell center region is screened out; the red blood cell center region and the adjacent suspected cytoplasm region are combined as a first blood cell region, and other suspected cytoplasm regions are second blood cell regions; according to the edge length and gray value of the first blood cell region and the second blood cell region, the cell aggregation degree of each blood cell image is obtained.

[0024] In the process of making blood smears, the thickness of the blood film changes from thick to shallow when the slide is pushed from the blood drop end to the other end. The blood film is too thick, which is easy to cause cell adhesion, and the blood film is too shallow, which is easy to cause misjudgment. And it is difficult to ensure that the blood drop amount and the pushing force of each smear are completely consistent when making, and the thickness of the blood film formed has differences. It is necessary to determine whether there is a uniform and dispersed region of blood cells that is easy to observe according to the characteristics of the blood smear itself. In the high-quality blood film region, the blood cells are uniformly distributed, and the cell aggregation and adhesion are less, and the cell morphology can be clearly observed. On the contrary, in other regions, the blood cells are not uniformly distributed. Because the blood drop is spread to both sides under the shearing force of the cover glass, the uniform distribution is that the two sides are slightly higher than the middle. However, if the thickness of the blood film is too high, the red blood cells in the middle are more aggregated than on both sides, so the cell uniformity distribution degree of the blood film at this position is lower, and it is not suitable to become a high-quality observation region. Because each blood cell image can be regarded as a local region in the smear, in order to evaluate the observable high-quality region, the cell aggregation degree in each blood cell image needs to be obtained first, and then the distribution change of the cell aggregation degree in the vertical direction is analyzed to determine whether it belongs to the case that the high-quality region is represented by the two sides being higher than the middle.

[0025] Firstly, considering that there is a clear gray difference between the background region and the cell region of the smear, the threshold segmentation algorithm can be used to segment the blood cell image. However, in the blood cell, the red blood cell has a special shape, the middle of the red blood cell is concave, and the periphery is convex, which causes the center of the red blood cell region to be a low gray region. In the threshold segmentation process, the center region will be identified as a background region, thereby causing the loss of gray information of the center region in the subsequent analysis process. Therefore, the suspected blood cell nucleus region and the suspected cytoplasm region are divided after threshold segmentation in the embodiment of the application. The suspected blood cell nucleus region is a connected domain formed by background information, and the suspected cytoplasm region is a connected domain formed by foreground information. If the suspected blood cell nucleus region is the red blood cell center region, it will show that the region is small and the roundness is high. Therefore, the red blood cell center region can be screened out according to the size of the suspected nucleus region and the edge curvature.

[0026] Preferably, in the embodiment of the present application, the blood cell image is processed by using a threshold segmentation algorithm to obtain a binary image; in the binary image, a connected domain composed of a type of pixel points with the highest gray scale is a suspected nucleus region, and a connected domain composed of another type of pixel points is a suspected cytoplasm region. Please refer to Figure 2 which shows a blood cell image gray scale diagram provided by an embodiment of the present application, and please refer to Figure 3 which shows a binary image provided by an embodiment of the present application, Figure 2 Figure 2 and Figure 3 It can be seen that Figure 3 the suspected cytoplasm region formed by the black foreground region in the binary image is a complete cell region or a region containing only cytoplasm, and the connected domain formed by the white background region is a red blood cell center region or a background region.

[0027] It should be noted that the threshold segmentation algorithm in the embodiment of the present application is the Otsu threshold algorithm, which is an automatic threshold segmentation algorithm, and the binary image can be obtained by processing the gray scale image of the blood cell image. The suspected nucleus region and the suspected cytoplasm region are determined by a connected domain algorithm, and the two pass algorithm can be used for extraction.

[0028] Preferably, the screening method of the red blood cell center region in the embodiment of the present application comprises: suspected nucleus regions smaller than a preset area threshold are regarded as second suspected nucleus regions, the curvatures of the second suspected nucleus regions are obtained and compared with the curvature of a standard circle, and the second suspected nucleus regions with a curvature error within a preset error range of the curvature of the standard circle are regarded as the red blood cell center regions.

[0029] It should be noted that the size of the area threshold is related to the magnification of the microscope. The larger the magnification of the microscope, the larger the shape of the cells in the image, and the more the pixel points of the red blood cell center region, i.e. the larger the area. In the embodiment of the present application, the microscope magnification used is 100 times, and the area threshold is set to 16, i.e. 16 pixel points, under the condition that the image resolution is 1920x1080. In the embodiment of the present application, the error range of the curvature is set to 10%, and the circular fitting can be realized by the least square method, which is a technology familiar to those skilled in the art and will not be described here.

[0030] After the red blood cell center region is determined, the red blood cell center region and the adjacent suspected cytoplasm region can be combined as a first blood cell region, i.e. a single mature red blood cell region or a superimposed cell region containing a mature red blood cell region. Other suspected cytoplasm regions are regarded as second blood cell regions, and the second blood cell regions are superimposed regions or single regions of other blood cells. ​

[0031] In a local area, if the cell aggregation degree is high, the obtained blood cell area presents a plurality of cell superposition conditions, resulting in a larger blood cell area; and because of the superposition, the area has poor light transmittance, is dark, and presents a clear gray value. Therefore, the cell aggregation degree of each blood cell image can be obtained according to the edge length and the gray value of the first blood cell area and the second blood cell area. That is, the larger the edge length, the larger the blood cell area, the smaller the gray value, the more the blood cell area presents the superposition characteristics, and the larger the cell aggregation degree.

[0032] Preferably, in the embodiment of the present application, the method for obtaining the cell aggregation degree comprises: The first blood cell area and the second blood cell area are taken as the analysis area, and the gray mean value of the analysis area in the blood cell image is obtained. Because the smaller the gray mean value, the darker the area, and the larger the cell superposition degree, the gray mean value is negatively correlated after mapping and multiplied by the edge length of the analysis area to obtain the initial aggregation degree of each analysis area. The larger the initial aggregation degree, the more the analysis area presents a plurality of cell superposition conditions. Because there are a plurality of analysis areas in the blood cell image, the initial aggregation degrees of all the analysis areas in the blood cell image are added and taken as the cell aggregation degree.

[0033] In the embodiment of the present application, the negative correlation mapping method adopts the form of reciprocal. Because the gray mean value of the analysis area may be 0, the reciprocal of the gray mean value plus a positive integer 1 is taken as the negative correlation mapping result.

[0034] Step S3: The cell aggregation degrees in each image set form a cell aggregation degree sequence according to the position sequence of the blood cell image; the cell aggregation degree uniformity of the cell aggregation degree sequence is obtained; the observation difficulty of each image set is obtained according to the cell aggregation degree uniformity and the cell aggregation degree; and the observable area in the scanning area is screened according to the observation difficulty.

[0035] Please refer to Figure 4It shows a schematic diagram of blood diffusion direction in a blood smearing process provided by an embodiment of the present application. During the movement of the pushing piece, the blood drop is subjected to the shearing force of the cover glass and the slide glass, spreads from the center to the two sides to form a blood film with a certain width. Since the shearing force on the two sides is usually greater than that on the center, the cell aggregation degree on the two sides is generally slightly higher than that on the center. However, when the blood drop is too large or the pushing piece speed is not appropriate, the blood concentration in the center is high, and the uniformity of cell distribution is poor. Therefore, the embodiment of the present application aims to find the area with the cell aggregation degree of the two sides being higher than that of the center as the observable area, and the area with the cell aggregation degree of the center being higher and the cell aggregation degree of the two sides being lower is difficult to observe. Therefore, the embodiment of the present application forms a cell aggregation degree sequence according to the position sequence of the blood cell image in each image set. The sequence is a sequence formed by the cell aggregation degree of the vertical local area. If the sequence shows the characteristics that the center element is larger and the two side elements are smaller, it means that the vertical area is difficult to observe. The cell aggregation degree uniformity of the cell aggregation degree sequence is obtained, and the observation difficulty of each image set is obtained according to the cell aggregation degree uniformity and the cell aggregation degree. That is, in a vertical area, the more the cell aggregation degree uniformity shows the characteristics that the center element is larger and the two side elements are smaller, and the larger the overall cell aggregation degree is, the higher the observation difficulty of the vertical area is. According to the observation difficulty, the observable area can be screened out, that is, the observable area is the merged area of the vertical areas with small observation difficulty.

[0036] Preferably, in the embodiment of the present application, the method for obtaining the cell aggregation degree uniformity comprises: According to the element value of the center position of the cell aggregation degree sequence, a contrast element value is obtained. The cell aggregation degree difference between the contrast element value and each element in the cell aggregation degree sequence is obtained. After the cell aggregation degree difference of all elements is negatively correlated and mapped, the cell aggregation degree uniformity is obtained. It should be noted that if the number of elements in the cell aggregation degree sequence is odd, the middle element value can be directly taken as the contrast element value. If it is even, the average value of the middle two element values is taken as the contrast element value. The smaller the cell aggregation degree difference between the contrast element value and other elements is, the more it shows that the cell aggregation degree on the two sides is relatively small compared with the cell aggregation degree on the center position, that is, it does not belong to the distribution characteristics of the observable area. Therefore, after being negatively correlated and mapped, the cell aggregation degree uniformity is obtained. The greater the cell aggregation degree uniformity is, the greater the observation difficulty is.

[0037] In the embodiment of the present application, the negative value of the difference of the cell aggregation degree is considered, and the difference is mapped by an exponential function with a natural constant as the base, i.e., the difference is taken as the power of the exponential function, and the mapping result of the function is a positive number. Similar to the negative correlation mapping of the average gray value, the inverse form can also be used for the negative correlation mapping, and details are not described herein.

[0038] Preferably, in the embodiment of the present application, the method for obtaining the observation difficulty comprises: normalizing the product of the cell aggregation degree uniformity and the average cell aggregation degree in the image set to obtain the observation difficulty.

[0039] The normalization algorithm in the embodiment of the present application can be realized by using range standardization, and in other embodiments of the present application, a function mapping method such as a hyperbolic tangent function can also be used, which is a technical means familiar to those skilled in the art, and details are not described herein. In the embodiment of the present application, the threshold interval is set to 0.33 to 0.45, the image set in which the observation difficulty is in the threshold interval is a local area of the observable region, the continuous local areas are merged, and the largest merged area is taken as the observable region. Please refer to Figure 5 which shows a schematic diagram of a blood smear observable region provided by an embodiment of the present application, in which the cell morphology is complete, and the distribution is relatively dispersed and uniform.

[0040] Step S4: obtaining the blood smear quality according to the area of the observable region and the shape integrity of the blood cell region in the observable region.

[0041] For the observable region, the larger the area is, the higher the coating quality of the blood smear is; the higher the integrity of the blood cell region in the observable region is, the more the coating process conforms to the provisions and requirements, and does not affect the normal observation of cells. Therefore, the blood smear quality can be obtained according to the area of the observable region and the shape integrity of the blood cell region in the observable region.

[0042] Preferably, in the embodiment of the present application, the complete blood cell connected domain should present non-concave in the image, and if cell fragmentation occurs in the coating process, the connected domain of the fragmented cell will present a shape with uneven concave and convex, and therefore the method for obtaining the shape integrity of the blood cell region in the observable region in the embodiment of the present application comprises: taking the first blood cell region and the second blood cell region in the observable region as a to-be-analyzed region, performing convex hull detection on each to-be-analyzed region to obtain the convex hull edge of each to-be-analyzed region. The greater the difference between the convex hull edge and the actual edge of the region is, the more likely the corresponding cell is a fragmented cell, and the more incomplete the shape is.

[0043] The edge point in the convex hull edge which is not the edge of the region to be analyzed is taken as a difference edge point, a ratio between a number of the difference edge points and a number of the edge points of the region to be analyzed is taken as an incompleteness, and the incompleteness is negatively correlated to obtain an initial shape completeness of each region to be analyzed. The greater the initial shape completeness is, the more likely the blood cell corresponding to the region to be analyzed is a complete blood cell region.

[0044] The region to be analyzed with the initial shape completeness greater than a preset first completeness threshold is taken as a complete blood cell region, and a proportion of the complete blood cell region in all regions to be analyzed is taken as a shape completeness of the blood cell region in the observable region.

[0045] In the embodiment of the present application, because the value range of the incompleteness is between 0 and 1, the positive integer 1 and the difference value of the incompleteness are directly taken as the negatively correlated mapping result to obtain the initial shape completeness. The first completeness threshold is set to 0.9.

[0046] Further, in the embodiment of the present application, the method for obtaining the blood smear quality comprises: An area proportion of the observable region in the blood region is obtained, and if the area proportion is less than a preset area proportion threshold, it is determined that the blood smear quality is not up to standard. In the embodiment of the present application, the area proportion threshold is set to 0.2.

[0047] If the area proportion is not less than the preset area proportion threshold, a shape completeness of the blood cell region in the observable region is obtained, if the shape completeness is greater than a preset second completeness threshold, it is determined that the blood smear quality is up to standard, and if the shape completeness is not greater than the preset second completeness threshold, it is determined that the blood smear quality is not up to standard. In the embodiment of the present application, the second completeness threshold is set to 0.9.

[0048] To sum up, the present application adopts the traversal scanning mode to obtain the blood cell images of all regions of the blood smear, and constructs the image set in column units. By screening the red blood cell center region and determining the first blood cell region and the second blood cell region to analyze together, the cell aggregation degree of each blood cell image can be accurately and effectively quantified. The observation difficulty of each image set is obtained by constructing the cell aggregation degree sequence and analyzing the uniformity thereof, and the observable region can be screened, so that the accurate blood smear quality evaluation is realized. The present application effectively quantifies the cell aggregation degree in the image by the image acquisition mode of traversal scanning, divides the correct observable region by the special morphology of the blood smear, and then realizes the accurate blood smear quality evaluation.

[0049] Based on the same inventive concept, the present application further provides a blood smear evaluation device based on medical examination, which comprises: The image acquisition module is used for image scanning in a direction perpendicular to the tail-to-head direction of the blood smear, and each blood cell image scanned in the perpendicular direction forms an image set in sequence during the scanning process; all image sets of the blood region on the blood smear are obtained; The cell aggregation degree acquisition module is used for obtaining a suspected blood cell nucleus region and a suspected cytoplasm region in the blood cell image by a threshold segmentation algorithm; a red blood cell center region is screened out according to the size and the edge curvature of the suspected blood cell nucleus region, and the red blood cell center region and the adjacent suspected cytoplasm region are combined as a blood cell region; the cell aggregation degree of each blood cell image is obtained according to the edge length and the gray value of the blood cell region; The observable region positioning module is used for forming a cell aggregation degree sequence according to the position sequence of the blood cell image according to the cell aggregation degree of each image set; the cell aggregation degree uniformity of the cell aggregation degree sequence is obtained; the observation difficulty of each image set is obtained according to the cell aggregation degree uniformity and the cell aggregation degree, and the observable region is screened out in the scanning region according to the observation difficulty. The blood smear quality evaluation module is used for obtaining the blood smear quality according to the area of the observable region and the shape integrity of the blood cell region in the observable region.

[0050] The application further provides a blood smear evaluation system based on medical examination, which comprises a memory, a processor, and a computer program stored in the memory and capable of running on the processor, and the processor implements the steps of any one of the blood smear evaluation methods based on medical examination.

[0051] It should be noted that the above-mentioned sequence of the embodiments of the application is only for description, and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or can be advantageous.

[0052] Each embodiment in the specification is described in a progressive manner, and the same and similar parts between each embodiment can be referred to each other, and each embodiment mainly describes the difference from other embodiments.

Claims

1. A blood smear evaluation method based on medical examination, characterized in that, The method comprises: Scanning the blood smear vertically and from the tail to the head, wherein the blood cell images obtained in each vertical scan during the scanning process sequentially constitute an image set; obtaining a set of all images of the blood area on the blood smear; A threshold segmentation algorithm is used to obtain suspected blood cell nucleus regions and suspected cytoplasm regions in a blood cell image; a red blood cell central region is screened out based on the size and edge curvature of the suspected cell nucleus region; the red blood cell central region and the adjacent suspected cytoplasm region are merged as a first blood cell region, and the other suspected cytoplasm regions are second blood cell regions; and the cell aggregation degree of each blood cell image is obtained based on the edge length and grayscale value of the first blood cell region and the second blood cell region; The cell aggregation degree in each image set forms a cell aggregation degree sequence according to the position sequence of the blood cell images; obtaining the cell aggregation degree uniformity of the cell aggregation degree sequence; obtaining the observation difficulty of each image set according to the cell aggregation degree uniformity and the cell aggregation degree, and screening an observable area in the scanning area according to the observation difficulty; The quality of the blood smear is obtained based on the area of ​​the observable region and the shape integrity of the blood cell region within the observable region.

2. A blood smear evaluation method based on medical examination according to claim 1, characterized in that: The obtaining of the suspected blood cell nucleus region and the suspected cytoplasm region in the blood cell image includes: The blood cell image is processed using a threshold segmentation algorithm to obtain a binary image; in the binary image, the connected domain composed of one type of pixel points with the highest grayscale is the suspected cell nucleus area, and the connected domain composed of another type of pixel points is the suspected cytoplasm area.

3. A blood smear evaluation method based on medical examination according to claim 1, characterized in that: The screening method for the central region of red blood cells comprises: The suspected cell nucleus region that is smaller than a preset area threshold is taken as the second suspected cell nucleus region, the curvature of the second suspected cell nucleus region is obtained and compared with the curvature of the standard circle, and the second suspected cell nucleus region whose curvature error with the standard circle is within a preset error range is taken as the red blood cell center region.

4. A blood smear evaluation method based on medical examination according to claim 1, characterized in that: The method for obtaining the degree of cell aggregation includes: The first blood cell region and the second blood cell region are used as the regions to be analyzed, and the grayscale mean of the regions to be analyzed in the blood cell image is obtained. The grayscale mean is negatively correlated with the edge length of the region to be analyzed and multiplied to obtain the initial aggregation degree of each region to be analyzed. The initial aggregation degrees of all the regions to be analyzed in the blood cell image are accumulated and used as the cell aggregation degree.

5. A blood smear evaluation method based on medical examination according to claim 1, characterized in that: The method for obtaining the uniformity of cell aggregation comprises: The comparison element value is obtained according to the element value at the center position of the cell aggregation degree sequence, and the cell aggregation degree difference between the comparison element value and each element in the cell aggregation degree sequence is obtained. The cell aggregation degree differences of all elements are negatively correlated and then accumulated to obtain the cell aggregation degree uniformity.

6. A blood smear evaluation method based on medical examination according to claim 1, characterized in that: The method for obtaining the observation difficulty includes: The observation difficulty is obtained by normalizing the product of the cell aggregation uniformity and the average cell aggregation degree in the image set.

7. A blood smear evaluation method based on medical examination according to claim 1, characterized in that: The method for obtaining the shape integrity of the blood cell region within the observable area includes: The first blood cell region and the second blood cell region in the observable area are used as the areas to be analyzed, and a convex hull detection is performed on each area to be analyzed to obtain the convex hull edge of each area to be analyzed. Edge points on the convex hull edge that are not on the edge of the area to be analyzed are used as difference edge points. The ratio between the number of difference edge points and the number of edge points on the edge of the area to be analyzed is used as the incompleteness. The incompleteness is negatively correlated with each other to obtain the initial shape completeness of each area to be analyzed. The areas to be analyzed whose initial shape completeness is greater than a preset first completeness threshold are used as complete blood cell areas, and the proportion of the number of complete blood cell areas in all areas to be analyzed is used as the shape completeness of the blood cell areas in the observable area.

8. A blood smear evaluation method based on medical examination according to claim 7, characterized in that: The method for obtaining the blood smear quality comprises: Obtaining an area ratio of the observable area to the blood area; if the area ratio is less than a preset area ratio threshold, determining that the blood smear quality does not meet the standard; If the area ratio is not less than the preset area ratio threshold, the shape integrity of the blood cell area in the observable area is obtained; if the shape integrity is greater than the preset second integrity threshold, the blood smear quality is judged to be up to standard; if the shape integrity is not greater than the preset second integrity threshold, the blood smear quality is judged to be unsatisfactory.

9. A blood smear evaluation device based on medical examination, characterized in that: The device comprises: The image acquisition module is used to scan the blood smear vertically and from the tail to the head. During the scanning process, the blood cell images obtained in each vertical scan are sequentially formed into an image set; all image sets of the blood area on the blood smear are obtained; A cell aggregation degree acquisition module is used to facilitate the threshold segmentation algorithm to obtain suspected blood cell nucleus regions and suspected cytoplasm regions in the blood cell image; screen out the red blood cell center region based on the size and edge curvature of the suspected cell nucleus region, and merge the red blood cell center region and the adjacent suspected cytoplasm region as the blood cell region; and obtain the cell aggregation degree of each blood cell image based on the edge length and grayscale value of the blood cell region; An observable region positioning module is configured to construct a cell aggregation degree sequence based on the positional order of the blood cell images according to the cell aggregation degree in each image set; obtain the cell aggregation degree uniformity of the cell aggregation degree sequence; obtain the observation difficulty of each image combination based on the cell aggregation degree uniformity and the cell aggregation degree; and screen out observable regions in the scanning area based on the observation difficulty; The blood smear quality assessment module is used to obtain the blood smear quality according to the area of ​​the observable region and the shape integrity of the blood cell region in the observable region.

10. A blood smear evaluation system based on medical examination, 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 blood smear evaluation method based on medical examination according to any one of claims 1 to 9 are implemented.

Images