Gastric tumor cell image extraction method and system

Abstract

The present invention provides a method and system for extracting gastric tumor cell images, and relates to the technical field of tumor image processing. The method includes collecting normal historical image data of the stomach, extracting image features of the normal state of the stomach, and forming normal gastric image feature data; collecting historical image data of gastric tumors, and extracting gastric tumor image features in combination with normal gastric image feature data, and forming gastric tumor image feature data; acquiring a target gastric image, and combining the normal gastric image feature data and the gastric tumor image feature data, performing tumor recognition analysis, and forming target image tumor feature information. The method realizes more accurate and efficient tumor image information extraction by utilizing big data analysis of gastric tumor images, thereby ensuring the accuracy and efficiency of tumor image feature extraction, and providing reliable data reference for doctors' diagnosis and treatment.

Description

Technical Field

[0001] The present invention relates to the technical field of tumor image processing, and in particular to a method and system for extracting gastric tumor cell images. Background Art

[0002] With the development of medical technology, the treatment of tumors is becoming more and more advanced and effective. Of course, the identification of tumors is also becoming more and more advanced and accurate. Gastric tumors are common tumor diseases. Due to their special location, they cannot be directly observed. Therefore, image technology is used to collect and identify tumors through the collected image information.

[0003] However, most of the extraction of gastric tumor image information is currently done manually, and due to the different experiences of different people, there are errors and inaccuracies in the extraction of tumor image information, which makes it impossible to provide doctors with accurate reference data for diagnosis and identification. At present, feature analysis based on big data is becoming more and more mature, and the extraction of image feature information can be achieved by using big data.

[0004] Therefore, it is an urgent problem to be solved to design a gastric tumor cell image extraction method and system, which can achieve more accurate and efficient tumor image information extraction by utilizing big data analysis of gastric tumor images, ensure the accuracy and efficiency of tumor image feature extraction, and provide reliable data reference for doctors' diagnosis and treatment. Summary of the invention

[0005] The purpose of the present invention is to provide a method for extracting gastric tumor cell images, which extracts image feature information of the stomach under normal conditions by acquiring normal stomach image data, and simultaneously acquires gastric tumor image data and extracts image feature information for the tumor. On this basis, the image feature information under normal conditions and the tumor image feature information are used to further optimize the features of the tumor image feature information, thereby ensuring the rationality and accuracy of the tumor feature information and making the recognition of the tumor features on the target gastric image more accurate and efficient. On the one hand, comprehensive and accurate tumor feature information can be established through big data analysis, and on the other hand, compared with manual extraction of tumor feature information, the recognition of tumor feature information obtained with the help of big data is more efficient and accurate, providing important and accurate data reference for medical personnel's diagnosis and treatment.

[0006] The purpose of the present invention is also to provide a gastric tumor cell image extraction system, which can realize feature information extraction of normal gastric image data and gastric tumor image data through configuration, so as to complete the identification and provision of tumor feature information of the target gastric image, form an efficient and accurate gastric tumor image feature information extraction method, lay an important material foundation for the extraction of gastric tumor feature information, and ensure the smooth and efficient extraction of gastric tumor image features.

[0007] In a first aspect, the present invention provides a method for extracting gastric tumor cell images, comprising collecting normal historical gastric image data, performing image feature extraction of the normal gastric state, and forming normal gastric image feature data; collecting historical gastric tumor image data, and performing gastric tumor image feature extraction in combination with the normal gastric image feature data, and forming gastric tumor image feature data; acquiring a target gastric image, and performing tumor recognition analysis in combination with the normal gastric image feature data and the gastric tumor image feature data, and forming target image tumor feature information.

[0008] In the present invention, the method extracts the image feature information of the stomach under normal conditions by acquiring the normal image data of the stomach, and at the same time acquires the image data of the stomach tumor and extracts the image feature information for the tumor. On this basis, the image feature information under normal conditions and the tumor image feature information are used to further optimize the features of the tumor image feature information, so as to ensure the rationality and accuracy of the tumor feature information, and make the identification of the tumor features on the target stomach image more accurate and efficient. On the one hand, comprehensive and accurate tumor feature information can be established through big data analysis. On the other hand, compared with the manual extraction of tumor feature information, the identification of tumor feature information obtained with the help of big data is more efficient and accurate, which provides important and accurate data reference for the diagnosis and treatment of medical personnel.

[0009] As a possible implementation method, normal historical stomach image data is collected, image features of the normal state of the stomach are extracted, and normal stomach image feature data is formed, including: grayscale processing is performed on different normal historical stomach images in the normal historical stomach image data to form corresponding normal historical stomach grayscale images; mutually orthogonal image transverse recognition directions and image longitudinal recognition directions are set, and grayscale feature information extraction is performed in two directions on different normal historical stomach images to respectively form corresponding normal transverse grayscale feature information of the image and normal longitudinal grayscale feature information of the image; normal transverse grayscale feature information of the image and normal longitudinal grayscale feature information of the image corresponding to different normal historical stomach grayscale images are collected to perform normal feature analysis to form normal stomach image feature data.

[0010] In the present invention, when extracting the image feature information of the normal state of the stomach from the normal historical image data of the stomach, two aspects need to be considered. One aspect is that the extracted adjustment information needs to accurately reflect the state of the normal state of the stomach, and can form a clear and direct contrast with the image feature information of the presence of a tumor in the stomach. The second aspect is that the normal feature information extraction method adopted can accurately obtain the feature information, avoiding the deviation caused by the subsequent use of the normal feature information for judgment and extraction, resulting in inaccurate analysis and judgment. Considering that for image data, the image information of the tumor area in the image where the tumor exists is mainly reflected in the color change, and directly using the color information will cause a huge amount of information to be analyzed due to the huge amount of color data information. In essence, the recognition of the tumor on the image is to distinguish the difference in color between the tumor area and other areas. Therefore, in order to reduce the amount of information analyzed and to ensure the accuracy of the extracted feature information, the image data is grayed to use the gray value to simplify and reasonably and accurately reflect the color change of the tumor area and the normal area. Therefore, in order to ensure the comparability of the data, when extracting feature information from the normal historical image of the stomach, the normal historical image of the stomach is grayed and then extracted. In addition, how to extract grayscale feature information from grayscale images to accurately display the feature information of normal images is the key to extracting tumor feature information. Considering that the most obvious difference between tumor areas and normal areas is that the grayscale changes on the boundary are very strong, that is, the grayscale value will jump greatly at the tumor boundary. Therefore, when extracting features from normal images, the changes in grayscale values ​​are also extracted. Of course, in order to reasonably analyze the changes in grayscale values, the extraction must be performed in a certain direction or order to ensure the changes in the grayscale values ​​of adjacent pixels of normal image data. Here, it is considered that extraction in a single direction will cause errors in the extracted data due to the pixel deviation when the image is generated. Therefore, comprehensive analysis from two different directions can improve the accuracy of feature information extraction, so it is performed in two intersecting directions. It should be noted that the two intersecting directions can be determined by the direction of the coordinate system of the device extracting the photo, or based on the position characteristics of the stomach in the acquired image, so that the directions of different images are consistent when they are extracted, which is convenient for subsequent reasonable and accurate analysis and processing.

[0011] As a possible implementation method, mutually orthogonal image transverse recognition directions and image longitudinal recognition directions are set, and grayscale feature information in two directions is extracted on different normal historical stomach images to form corresponding normal transverse grayscale feature information of the image and normal longitudinal grayscale feature information of the image, respectively, including: setting a transverse direction extraction step length m, and extracting grayscale values ​​of all pixels in the corresponding column in the transverse recognition direction of the image for different normal historical stomach grayscale images with the transverse direction extraction step length m as an interval, and fitting is performed according to the grayscale value of the corresponding column along the longitudinal recognition direction of the image to form a normal grayscale value change function of the transverse column of the corresponding column. , where p represents the number of different normal historical grayscale images of the stomach, n represents the number of different columns sequentially extracted in the horizontal recognition direction of the normal historical grayscale image of the stomach numbered p, and a is a natural number and represents the pixel number in the horizontal column; the vertical direction extraction step i is set, and the grayscale values ​​of all pixels in the corresponding row are sequentially extracted in the vertical recognition direction of the image for different normal historical grayscale images of the stomach with the vertical direction extraction step i as an interval, and the grayscale values ​​of the corresponding row along the horizontal recognition direction of the image are fitted to form the vertical row normal grayscale value change function of the corresponding row , where k represents the numbers of different rows sequentially extracted in the longitudinal recognition direction of the normal historical grayscale image of the stomach numbered p, and b is a natural number and represents the pixel number in the longitudinal row.

[0012] In the present invention, for the extraction of normal image feature information in the horizontal recognition direction, considering that the grayscale value of the image is in pixels, if all are extracted in sequence, it will cause a huge amount of data, so a certain interval, that is, the horizontal extraction step length, can be set to extract at intervals. Of course, the size of the interval can be set as needed. For the pixel values ​​extracted at intervals, they are extracted in units of vertical columns at a certain horizontal interval position in the horizontal direction, so that continuous grayscale value change information on the vertical column can be formed. Here, it should be noted that the grayscale values ​​corresponding to all pixels on each horizontal column extracted are counted in units of pixels, which is a series of discrete data. By fitting to form a change function, two effects can be achieved. One is that the grayscale value error generated when the image is generated can be eliminated, that is, the effect of reducing the noise of the image grayscale value is achieved, so that the formed data is more accurate. On the other hand, the grayscale value change function obtained under the condition of trying to envelop the grayscale value of each pixel obtained on the image reasonably provides a value tolerance for each pixel, so that the grayscale value value on the pixel is more reasonable. Similarly, the grayscale feature information extraction in the longitudinal direction is also performed by using the step interval for interval extraction, and a reasonable fitting is performed on the grayscale values ​​of the pixels in each horizontal row formed by the extraction to ensure the accuracy and rationality of the grayscale value change function formed.

[0013] As a possible implementation method, normal transverse grayscale feature information and normal longitudinal grayscale feature information corresponding to different normal historical grayscale images of the stomach are collected to perform normal feature analysis to form normal stomach image feature data, including: determining the normal grayscale value change function of all transverse columns in different normal historical grayscale images of the stomach The normal gray value change rate function of the horizontal column ,in, for The first derivative of the normal gray value change rate function of different horizontal columns with the same arrangement number in different normal historical gray images of the stomach Delimit the range at each pixel point to form the normal grayscale change rate range data for each horizontal column ; Determine the normal grayscale value change function of all longitudinal rows in the normal historical grayscale images of different stomachs The vertical row normal gray value change rate function ,in, for The first derivative of the normal grayscale value change rate function of different vertical rows with the same arrangement number in different normal historical grayscale images of the stomach The range is defined at each pixel point to form the normal gray value change rate range data for each vertical row .

[0014] In the present invention, it should be noted that for stomach images with tumors, the change in grayscale value at the boundary of the tumor area is relatively large, and the change in grayscale value should be represented by the change rate of the grayscale value. Therefore, for feature extraction of normal image data, after obtaining the grayscale value change function, the change rate information can be extracted to form data that can be compared with subsequent tumor image features.

[0015] As a possible implementation method, historical image data of gastric tumors are collected, and gastric tumor image features are extracted in combination with normal gastric image feature data to form gastric tumor image feature data, including: grayscale processing is performed on different gastric tumor historical images in the gastric tumor historical image data to form corresponding gastric tumor historical grayscale images; grayscale feature information is extracted in the horizontal image recognition direction and the vertical image recognition direction of different gastric tumor historical grayscale images to form image tumor horizontal grayscale feature information and image tumor vertical grayscale feature information respectively; image tumor horizontal grayscale feature information and image tumor vertical grayscale feature information corresponding to different gastric tumor historical grayscale images are collected to perform tumor feature analysis to form preliminary gastric tumor image feature data; gastric tumor recognition feature extraction is performed in combination with normal gastric image feature data and gastric tumor image feature data to form preliminary gastric tumor image feature data.

[0016] In the present invention, for the historical image data of gastric tumors, corresponding to the historical image data of normal stomach, it is first necessary to extract the grayscale value change information in two directions, and then obtain accurate change rate data. Of course, only extracting the tumor image feature information cannot accurately use the tumor image feature information to extract tumor information. After all, in the preliminary extracted tumor image feature information, the irregularity of the tumor boundary may exist in the intersection of the normal image feature information. Therefore, in order to obtain the tumor feature information more accurately, after obtaining the preliminary tumor image feature information, it can be compared and analyzed with the normal stomach feature data to extract more accurate and reasonable gastric tumor image feature information.

[0017] As a possible implementation method, grayscale feature information of different gastric tumor historical grayscale images is extracted in the horizontal recognition direction and the vertical recognition direction of the image to form the horizontal grayscale feature information of the image tumor and the vertical grayscale feature information of the image tumor, respectively, including: for different gastric tumor historical grayscale images, grayscale values ​​of corresponding columns are extracted in sequence in the horizontal recognition direction of the image with a horizontal extraction step length m as an interval, and grayscale values ​​of corresponding columns along the vertical recognition direction of the image are fitted to form a horizontal column tumor grayscale value change function of the corresponding column , where q represents the number of different gastric tumor historical grayscale images, u represents the number of different columns sequentially extracted in the horizontal recognition direction of the gastric tumor historical grayscale image numbered q, and u=n; for different gastric tumor historical grayscale images, the grayscale values ​​of all pixels in the corresponding row are sequentially extracted in the vertical recognition direction of the image with the vertical extraction step i as the interval, and the grayscale values ​​of the corresponding row along the horizontal recognition direction of the image are fitted to form the longitudinal row tumor grayscale value change function of the corresponding row , where v represents the numbers of different rows sequentially extracted in the longitudinal recognition direction of the historical grayscale image of the gastric tumor numbered q, and v=k.

[0018] In the present invention, the extraction of grayscale value change information of historical grayscale images of gastric tumors is the same as the extraction method of grayscale value change information of normal gastric images. On the one hand, it can ensure the accuracy and rationality of the extracted feature information, and on the other hand, it can achieve good comparability with normal feature information, thereby realizing subsequent correct comparative analysis of normal feature information and tumor feature information.

[0019] As a possible implementation method, the image tumor transverse grayscale feature information and image tumor longitudinal grayscale feature information corresponding to different gastric tumor historical grayscale images are collected to perform tumor feature analysis to form preliminary feature data of gastric tumor images, including: determining the grayscale value change function of all transverse columns of tumors in different gastric tumor historical grayscale images The gray value change rate function of the horizontal column tumor ,in, for The first-order derivative of; the gray value change rate function of tumors in different transverse columns with the same arrangement number in different gastric tumor historical gray images The range is defined at each pixel point to form the range data of tumor grayscale change rate corresponding to each horizontal column. ; Determine the gray value change function of all longitudinal rows of tumors in different gastric tumor historical gray images The function of the change rate of the longitudinal tumor gray value ,in, for The first-order derivative of; the gray value change rate function of different longitudinal rows of tumors with the same arrangement number in different historical gray images of gastric tumors The range is delineated at each pixel point to form the range data of the gray value change rate of the longitudinal tumor corresponding to each longitudinal row. .

[0020] In the present invention, similarly to the extraction of normal stomach feature information, after the stomach tumor feature information is extracted into the grayscale value change function in two directions, it is necessary to obtain the corresponding grayscale value change rate function. The grayscale value change rate is the accurate data reflecting the grayscale value change of the tumor boundary.

[0021] As a possible implementation method, the normal stomach image feature data and the preliminary stomach tumor image feature data are combined to extract the stomach tumor recognition feature data to form the stomach tumor image feature data, including: for the horizontal columns with the same arrangement number, the normal grayscale change rate range data And the horizontal column tumor grayscale change rate range data , compare the change rate range of each pixel point a in turn, and determine the tumor grayscale change rate range data of the horizontal column at the pixel point The horizontal column tumor gray value range and the horizontal column normal gray change rate range data The horizontal column normal gray value range does not intersect the horizontal column tumor gray value range, mark the corresponding pixel point as the horizontal suspected tumor boundary point, and mark the corresponding horizontal column tumor gray value range as the horizontal column suspected tumor gray boundary change rate value range; for the vertical row normal gray value change rate range data with the same arrangement number And longitudinal row tumor gray value change rate range data , compare the change rate range of each pixel point b in turn, and determine the change rate range data of the tumor gray value in the longitudinal row at the pixel point The vertical row of tumor gray value range and vertical row of normal gray value change rate range data The vertical row of normal grayscale value ranges do not intersect the vertical row of tumor grayscale value ranges, mark the corresponding pixel points as longitudinal suspected tumor boundary points, and mark the corresponding vertical row of tumor grayscale value ranges as longitudinal row suspected tumor grayscale boundary change rate value ranges; determine the pixel position where the horizontal suspected tumor boundary point and the vertical suspected tumor boundary point are the same pixel point, and mark them as tumor boundary pixel positions; extract the horizontal column suspected tumor grayscale boundary change rate value range corresponding to the tumor boundary pixel position, and calibrate it as the horizontal column tumor grayscale boundary change rate value range; extract the vertical row suspected tumor grayscale boundary change rate value range corresponding to the tumor boundary pixel position, and calibrate it as the vertical row tumor grayscale boundary change rate value range; combine the horizontal column tumor grayscale boundary change rate value range and the vertical row tumor grayscale boundary change rate value range to form tumor boundary grayscale change rate range data; gather all tumor boundary grayscale change rate range data to form gastric tumor image feature data.

[0022] In the present invention, the feature data of the normal stomach image is compared and analyzed with the preliminary feature data of the stomach tumor image, mainly to optimize the feature data of the stomach tumor image, so as to provide more accurate reference comparison information when the tumor feature information of the stomach image is analyzed later. It can be understood that for the pixel points at the boundary of the tumor, whether in the horizontal direction or the vertical direction, a sudden change in the gray value change rate will occur, which is different from the normal gray value change rate. Therefore, after the gray value change rate in the horizontal direction and the vertical direction are compared and analyzed respectively, it is also necessary to determine whether the sudden change in the gray value change rate in the two directions occurs at the same pixel point. Only when the change rate in the two directions is changed can the corresponding pixel point be determined to be the tumor boundary point, and the value range of the gray value change rate in the two directions corresponding to this pixel point is the basis for truly and accurately judging the tumor boundary. Since the position of the tumor is different in different images and the area is also different, it is only necessary to extract the gray value change rate range as feature information, which can be used to judge the tumor boundary of the target stomach image.

[0023] As a possible implementation method, a target stomach image is obtained, and tumor recognition analysis is performed in combination with feature data of a normal stomach image and feature data of a stomach tumor image to form target image tumor feature information, including: grayscale processing is performed on the target stomach image to form a target stomach grayscale image; the target stomach grayscale image is extracted in sequence in the horizontal recognition direction of the image with a horizontal extraction step length m as an interval, the grayscale values ​​of all pixels in the corresponding column are extracted, and fitting is performed according to the grayscale value of the corresponding column along the vertical recognition direction of the image to form a horizontal column target grayscale value change function of the corresponding column. , and according to the gray value change function of different horizontal columns Determine the corresponding horizontal column target gray value change rate function ; Extract the grayscale values ​​of all pixels in the corresponding column in the target stomach grayscale image in the longitudinal direction of the image with the longitudinal extraction step length i as the interval, and fit the grayscale values ​​of the corresponding row along the horizontal recognition direction of the image to form the longitudinal row target grayscale value change function of the corresponding row , and according to the gray value change function of different vertical rows Determine the corresponding vertical row target gray value change rate function ; Determine the gray value change rate function of the horizontal column target And the vertical row target gray value change rate function The horizontal target grayscale value change rate and the vertical target grayscale value change rate on the same pixel point are calculated, and the corresponding pixel point is marked as the target analysis pixel point; according to the gastric tumor image feature data, the pixel points on the target gastric grayscale image are marked in the following manner: for any target analysis pixel point, if there is a tumor boundary grayscale change rate range data in the gastric tumor image recognition feature data that satisfies the horizontal target grayscale value change rate of the target analysis pixel point belongs to the corresponding horizontal column tumor grayscale boundary change rate value range, and the vertical target grayscale value change rate of the target analysis pixel point belongs to the corresponding vertical row tumor grayscale boundary change rate value range. range, the target analysis pixel point is calibrated as a suspected tumor boundary pixel point; the suspected tumor boundary pixel points that are adjacent in the horizontal recognition direction and the vertical recognition direction of the image are connected, and the following gastric tumor analysis is performed: if the adjacent suspected tumor boundary pixel points are connected to form a closed curve, the closed curve is calibrated for the tumor position; if the adjacent suspected tumor boundary pixel points are connected and form a closed curve with the image boundary, the closed area is calibrated for the tumor position; if the adjacent suspected tumor boundary pixel points are connected but do not form a closed curve, and cannot form a closed curve with the image boundary, the target gastric image is determined to be a normal image.

[0024] In the present invention, when extracting image feature information obtained by big data, the target stomach image is first grayed, and the gray value change rate function is extracted on the grayed image in two directions at the same interval as the feature information extraction of the normal image and the tumor image. Then, the gray value change rate in the transverse direction and the gray value change rate in the longitudinal direction of the intersecting pixel points are compared and analyzed with the gastric tumor image feature data. The gray value change rate values ​​in the two directions are correlated. It is necessary to ensure that the gray value change rate in the two directions of the intersecting pixel points belongs to the range defined by the combination of the change rate range in the same transverse and longitudinal directions in the gastric tumor image feature data before the pixel point can be identified as a tumor boundary point. Of course, the determined pixel points are discrete points, while the tumor boundary is continuous, so after determining all the tumor boundary pixel points, it is necessary to determine whether these pixel points can form a reasonable boundary. It can be understood that if the pixel points formed by the intersection in the transverse direction and the longitudinal direction are the boundaries of the tumor, then the pixel points are adjacent and can be connected. Considering the size of the extracted image area, it may not be able to completely cover the entire tumor area, so two situations are considered when judging. One is that the connection between adjacent pixels can form a closed curve, and the closed curve formed is the tumor area. The other is that the adjacent pixels can only form a curve, but this curve intersects with the boundary of the image and can form a closed curve with the boundary of the image. The area within the closed curve is the part of the tumor that appears in the image. It should be noted that the adjacency of pixel points needs to meet the adjacency in both the horizontal and vertical directions. Discontinuities in any direction cannot become the connection of a closed curve.

[0025] In a second aspect, the present invention provides a gastric tumor cell image extraction system, which is configured to: collect historical normal gastric image data, perform image feature extraction of the normal gastric state, and form normal gastric image feature data; collect historical gastric tumor image data, and perform gastric tumor image feature extraction in combination with the normal gastric image feature data, and form gastric tumor image feature data; obtain a target gastric image, and perform tumor recognition analysis in combination with the normal gastric image feature data and the gastric tumor image feature data, and form target image tumor feature information.

[0026] In the present invention, the system is configured to extract feature information from normal stomach image data and stomach tumor image data to provide identification of tumor feature information of the target stomach image, thereby forming an efficient and accurate method for extracting feature information from stomach tumor images, laying an important material foundation for the extraction of stomach tumor feature information and ensuring the smooth and efficient extraction of stomach tumor image features.

[0027] The beneficial effects of a gastric tumor cell image extraction method and system provided by the present invention are as follows:

[0028] This method extracts the image feature information of the stomach under normal conditions by acquiring normal stomach image data, and at the same time acquires stomach tumor image data and extracts image feature information for the tumor. On this basis, the image feature information under normal conditions and tumor image feature information are used to further optimize the tumor image feature information, ensuring the rationality and accuracy of the tumor feature information, making the recognition of tumor features on the target stomach image more accurate and efficient. On the one hand, comprehensive and accurate tumor feature information can be established through big data analysis. On the other hand, compared with manual extraction of tumor feature information, identification with the help of tumor feature information obtained by big data is more efficient and accurate, providing important and accurate data reference for medical personnel's diagnosis and treatment.

[0029] The system is configured to extract feature information from normal stomach image data and stomach tumor image data to provide identification of tumor feature information of target stomach images, thereby forming an efficient and accurate method for extracting feature information from stomach tumor images, laying an important material foundation for the extraction of feature information from stomach tumors, and ensuring the smooth and efficient extraction of feature information from stomach tumor images. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments of the present invention are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0031] Figure 1 A step diagram of a method for extracting gastric tumor cell images provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0032] The technical solutions in the embodiments of the present invention will be described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0033] refer to Figure 1The embodiment of the present invention provides a method for extracting gastric tumor cell images. The method extracts the image feature information of the stomach under normal conditions by acquiring normal stomach image data, and simultaneously acquires gastric tumor image data and extracts image feature information for the tumor. On this basis, the image feature information under normal conditions and the tumor image feature information are used to further optimize the features of the tumor image feature information, thereby ensuring the rationality and accuracy of the tumor feature information and making the identification of the tumor features on the target stomach image more accurate and efficient. On the one hand, comprehensive and accurate tumor feature information can be established through big data analysis. On the other hand, compared with manual extraction of tumor feature information, identification with the help of tumor feature information obtained with the help of big data is more efficient and accurate, providing important and accurate data reference for medical personnel's diagnosis and treatment.

[0034] The gastric tumor cell image extraction method specifically comprises the following steps:

[0035] S1: Collecting historical normal stomach image data, extracting image features of the normal state of the stomach, and forming normal stomach image feature data.

[0036] Collecting normal historical image data of the stomach, extracting image features of the normal state of the stomach, and forming normal image feature data of the stomach, including: performing grayscale processing on different normal historical stomach images in the normal historical stomach image data to form corresponding normal historical stomach grayscale images; setting mutually orthogonal image transverse recognition directions and image longitudinal recognition directions, and extracting grayscale feature information in two directions on different normal historical stomach images to respectively form corresponding normal transverse grayscale feature information of the image and normal longitudinal grayscale feature information of the image; and collecting normal transverse grayscale feature information of the image and normal longitudinal grayscale feature information of the image corresponding to different normal historical stomach grayscale images, performing normal feature analysis, and forming normal image feature data of the stomach.

[0037] When extracting the image feature information of the normal state of the stomach from the normal historical image data of the stomach, two aspects need to be considered. One aspect is that the extracted adjustment information needs to accurately reflect the state of the normal state of the stomach, and can form a clear and direct contrast with the image feature information of the stomach with tumors. The second aspect is that the normal feature information extraction method can accurately obtain the feature information, avoiding the deviation caused by the subsequent use of normal feature information for judgment and extraction, resulting in inaccurate analysis and judgment. Considering that for image data, the image information of the tumor area in the image where the tumor exists is mainly reflected in the color change, and the direct use of color information will cause a huge amount of information to be analyzed due to the huge amount of color data information. In essence, the recognition of tumors in the image is to distinguish the difference in color between the tumor area and other areas. Therefore, in order to reduce the amount of information analyzed and to ensure the accuracy of the extracted feature information, the image data is grayed to use gray values ​​to simplify and reasonably and accurately reflect the color changes of the tumor area and the normal area. Therefore, in order to ensure the comparability of the data, when extracting feature information from the normal historical images of the stomach, the normal historical images of the stomach are grayed and then extracted. In addition, how to extract grayscale feature information from grayscale images to accurately display the feature information of normal images is the key to extracting tumor feature information. Considering that the most obvious difference between tumor areas and normal areas is that the grayscale changes on the boundary are very strong, that is, the grayscale value will jump greatly at the tumor boundary. Therefore, when extracting features from normal images, the changes in grayscale values ​​are also extracted. Of course, in order to reasonably analyze the changes in grayscale values, the extraction must be performed in a certain direction or order to ensure the changes in the grayscale values ​​of adjacent pixels of normal image data. Here, it is considered that extraction in a single direction will cause errors in the extracted data due to the pixel deviation when the image is generated. Therefore, comprehensive analysis from two different directions can improve the accuracy of feature information extraction, so it is performed in two intersecting directions. It should be noted that the two intersecting directions can be determined by the direction of the coordinate system of the device extracting the photo, or based on the position characteristics of the stomach in the acquired image, so that the directions of different images are consistent when they are extracted, which is convenient for subsequent reasonable and accurate analysis and processing.

[0038] The method comprises setting mutually orthogonal image transverse recognition directions and image longitudinal recognition directions, and extracting grayscale feature information in two directions on different normal historical stomach images to form corresponding normal transverse grayscale feature information and normal longitudinal grayscale feature information of the image, respectively, including: setting a transverse direction extraction step length m, and extracting grayscale values ​​of all pixels in the corresponding column in the transverse recognition direction of the image with the transverse direction extraction step length m as an interval for different normal historical stomach grayscale images, and fitting the grayscale values ​​of the corresponding column along the longitudinal recognition direction of the image to form a normal grayscale value change function of the transverse column of the corresponding column. , where p represents the number of different normal historical grayscale images of the stomach, n represents the number of different columns sequentially extracted in the horizontal recognition direction of the normal historical grayscale image of the stomach numbered p, and a is a natural number and represents the pixel number in the horizontal column; the vertical direction extraction step i is set, and the grayscale values ​​of all pixels in the corresponding row are sequentially extracted in the vertical recognition direction of the image for different normal historical grayscale images of the stomach with the vertical direction extraction step i as an interval, and the grayscale values ​​of the corresponding row along the horizontal recognition direction of the image are fitted to form the vertical row normal grayscale value change function of the corresponding row , where k represents the numbers of different rows sequentially extracted in the longitudinal recognition direction of the normal historical grayscale image of the stomach numbered p, and b is a natural number and represents the pixel number in the longitudinal row.

[0039] For the extraction of normal image feature information in the horizontal recognition direction, considering that the grayscale value of the image is in pixels, if all are extracted in sequence, it will cause a huge amount of data, so a certain interval, that is, the horizontal extraction step length, can be set to extract at intervals. Of course, the size of the interval can be set as needed. For the pixel values ​​extracted at intervals, they are extracted in units of vertical columns at a certain horizontal interval position in the horizontal direction, so that continuous grayscale value change information on the vertical column can be formed. Here, it should be noted that the grayscale values ​​corresponding to all pixels on each horizontal column extracted are counted in units of pixels, which is a series of discrete data. By fitting to form a change function, two effects can be achieved. One is that the grayscale value error generated when the image is generated can be eliminated, that is, the effect of reducing the noise of the image grayscale value can be achieved, so that the formed data is more accurate. On the other hand, the grayscale value change function obtained under the condition of trying to envelop the grayscale value of each pixel obtained on the image reasonably provides a value tolerance for each pixel, so that the grayscale value of the pixel is more reasonable. Similarly, the grayscale feature information extraction in the longitudinal direction is also performed by using the step interval for interval extraction, and a reasonable fitting is performed on the grayscale values ​​of the pixels in each horizontal row formed by the extraction to ensure the accuracy and rationality of the grayscale value change function formed.

[0040] The normal transverse grayscale feature information and the normal longitudinal grayscale feature information of the image corresponding to different normal historical grayscale images of the stomach are collected to perform normal feature analysis to form normal stomach image feature data, including: determining the normal grayscale value change function of all transverse columns in different normal historical grayscale images of the stomach The normal gray value change rate function of the horizontal column ,in, for The first derivative of the normal gray value change rate function of different horizontal columns with the same arrangement number in different normal historical gray images of the stomach Delimit the range at each pixel point to form the normal grayscale change rate range data for each horizontal column ; Determine the normal grayscale value change function of all longitudinal rows in the normal historical grayscale images of different stomachs The vertical row normal gray value change rate function ,in, for The first derivative of the normal grayscale value change rate function of different vertical rows with the same arrangement number in different normal historical grayscale images of the stomach The range is defined at each pixel point to form the normal gray value change rate range data for each vertical row .

[0041] It should be noted that for stomach images with tumors, the change in grayscale value at the boundary of the tumor area is relatively large, and the change in grayscale value should be represented by the grayscale value change rate. Therefore, for feature extraction of normal image data, after obtaining the grayscale value change function, the change rate information can be extracted to form data that can be compared with subsequent tumor image features.

[0042] S2: Collect historical image data of gastric tumors, and extract gastric tumor image features in combination with normal gastric image feature data to form gastric tumor image feature data.

[0043] Historical image data of gastric tumors are collected, and gastric tumor image features are extracted in combination with normal gastric image feature data to form gastric tumor image feature data, including: grayscale processing is performed on different gastric tumor historical images in the gastric tumor historical image data to form corresponding gastric tumor historical grayscale images; grayscale feature information is extracted in the horizontal image recognition direction and the vertical image recognition direction of different gastric tumor historical grayscale images to form image tumor horizontal grayscale feature information and image tumor vertical grayscale feature information respectively; image tumor horizontal grayscale feature information and image tumor vertical grayscale feature information corresponding to different gastric tumor historical grayscale images are collected to perform tumor feature analysis to form preliminary gastric tumor image feature data; gastric tumor recognition feature is extracted in combination with normal gastric image feature data and gastric tumor image feature data to form preliminary gastric tumor image feature data.

[0044] For the historical image data of gastric tumors, corresponding to the historical image data of normal stomach, it is first necessary to extract the grayscale value change information in two directions, and then obtain accurate change rate data. Of course, simply extracting the tumor image feature information cannot accurately use the tumor image feature information to extract tumor information. After all, in the preliminary extracted tumor image feature information, the irregularity of the tumor boundary may exist in the intersection of the normal image feature information. Therefore, in order to obtain the tumor feature information more accurately, it is possible to compare and analyze the preliminary tumor image feature information with the normal stomach feature data after obtaining it, and extract more accurate and reasonable gastric tumor image feature information.

[0045] Extracting grayscale feature information in the horizontal recognition direction and the vertical recognition direction of different gastric tumor historical grayscale images to form image tumor horizontal grayscale feature information and image tumor vertical grayscale feature information respectively, including: extracting grayscale values ​​of corresponding columns in the horizontal recognition direction of the image with a horizontal direction extraction step length m as an interval, and fitting according to the grayscale values ​​of the corresponding columns along the vertical recognition direction of the image to form a horizontal column tumor grayscale value change function of the corresponding column , where q represents the number of different gastric tumor historical grayscale images, u represents the number of different columns sequentially extracted in the horizontal recognition direction of the gastric tumor historical grayscale image numbered q, and u=n; for different gastric tumor historical grayscale images, the grayscale values ​​of all pixels in the corresponding row are sequentially extracted in the vertical recognition direction of the image with the vertical extraction step i as the interval, and the grayscale values ​​of the corresponding row along the horizontal recognition direction of the image are fitted to form the longitudinal row tumor grayscale value change function of the corresponding row , where v represents the numbers of different rows sequentially extracted in the longitudinal recognition direction of the historical grayscale image of the gastric tumor numbered q, and v=k.

[0046] The extraction of grayscale value change information of historical grayscale images of gastric tumors is the same as the extraction method of grayscale value change information of normal gastric images. On the one hand, it can ensure the accuracy and rationality of the extracted feature information, and on the other hand, it can achieve good comparability with normal feature information, thereby realizing the subsequent correct comparative analysis of normal feature information and tumor feature information.

[0047] Collect the image tumor transverse grayscale feature information and image tumor longitudinal grayscale feature information corresponding to different gastric tumor historical grayscale images, perform tumor feature analysis, and form preliminary feature data of gastric tumor images, including: determining the grayscale value change function of all transverse columns of different gastric tumor historical grayscale images The gray value change rate function of the horizontal column tumor ,in, for The first-order derivative of; the gray value change rate function of tumors in different transverse columns with the same arrangement number in different gastric tumor historical gray images The range is defined at each pixel point to form the range data of tumor grayscale change rate corresponding to each horizontal column. ; Determine the gray value change function of all longitudinal rows of tumors in different gastric tumor historical gray images The function of the change rate of the longitudinal tumor gray value ,in, for The first-order derivative of; the gray value change rate function of different longitudinal rows of tumors with the same arrangement number in different historical gray images of gastric tumors The range is delineated at each pixel point to form the range data of the gray value change rate of the longitudinal tumor corresponding to each longitudinal row. .

[0048] Similarly, just like the extraction of normal stomach feature information, after the stomach tumor feature information is extracted into the grayscale value change function in two directions, it is necessary to obtain the corresponding grayscale value change rate function. The grayscale value change rate is the accurate data reflecting the grayscale value change of the tumor boundary.

[0049] Combine the normal stomach image feature data and the preliminary stomach tumor image feature data to extract stomach tumor recognition features and form stomach tumor image feature data, including: normal grayscale change rate range data for horizontal columns with the same arrangement number And the horizontal column tumor grayscale change rate range data , compare the change rate range of each pixel point a in turn, and determine the tumor grayscale change rate range data of the horizontal column at the pixel point The horizontal column tumor gray value range and the horizontal column normal gray change rate range data The horizontal column normal gray value range does not intersect the horizontal column tumor gray value range, mark the corresponding pixel point as the horizontal suspected tumor boundary point, and mark the corresponding horizontal column tumor gray value range as the horizontal column suspected tumor gray boundary change rate value range; for the vertical row normal gray value change rate range data with the same arrangement number And longitudinal row tumor gray value change rate range data , compare the change rate range of each pixel point b in turn, and determine the change rate range data of the tumor gray value in the longitudinal row at the pixel point The vertical row of tumor gray value range and vertical row of normal gray value change rate range data The vertical row of normal grayscale value ranges do not intersect the vertical row of tumor grayscale value ranges, mark the corresponding pixel points as longitudinal suspected tumor boundary points, and mark the corresponding vertical row of tumor grayscale value ranges as longitudinal row suspected tumor grayscale boundary change rate value ranges; determine the pixel position where the horizontal suspected tumor boundary point and the vertical suspected tumor boundary point are the same pixel point, and mark them as tumor boundary pixel positions; extract the horizontal column suspected tumor grayscale boundary change rate value range corresponding to the tumor boundary pixel position, and calibrate it as the horizontal column tumor grayscale boundary change rate value range; extract the vertical row suspected tumor grayscale boundary change rate value range corresponding to the tumor boundary pixel position, and calibrate it as the vertical row tumor grayscale boundary change rate value range; combine the horizontal column tumor grayscale boundary change rate value range and the vertical row tumor grayscale boundary change rate value range to form tumor boundary grayscale change rate range data; gather all tumor boundary grayscale change rate range data to form gastric tumor image feature data.

[0050] The comparative analysis of the feature data of the normal stomach image and the preliminary feature data of the stomach tumor image is mainly to optimize the feature data of the stomach tumor image, so as to provide more accurate reference comparison information for the subsequent analysis of the tumor feature information of the stomach image. It can be understood that for the pixel points at the boundary of the tumor, whether in the horizontal direction or the vertical direction, there will be a sudden change in the gray value change rate that is different from the normal gray value change rate. Therefore, after the gray value change rate in the horizontal direction and the vertical direction are compared and analyzed respectively, it is also necessary to determine whether the sudden change in the gray value change rate in the two directions occurs at the same pixel point. Only when the change rate in both directions is changed can the corresponding pixel point be determined to be the tumor boundary point, and the value range of the gray value change rate in the two directions corresponding to this pixel point is the basis for truly and accurately judging the tumor boundary. Since the location of the tumor is different and the area is different in different images, it is only necessary to extract the gray value change rate range as feature information, which can be used to judge the tumor boundary of the target stomach image.

[0051] S3: Acquire a target stomach image, and perform tumor recognition analysis based on the normal stomach image feature data and the stomach tumor image feature data to form target image tumor feature information.

[0052] A target stomach image is acquired, and tumor recognition analysis is performed in combination with feature data of a normal stomach image and feature data of a stomach tumor image to form target image tumor feature information, including: grayscale processing is performed on the target stomach image to form a target stomach grayscale image; the target stomach grayscale image is extracted in sequence in a horizontal recognition direction of the image with a horizontal extraction step length m as an interval, the grayscale values ​​of all pixels in a corresponding column are extracted, and fitting is performed according to the grayscale value of the corresponding column along the vertical recognition direction of the image to form a horizontal column target grayscale value change function of the corresponding column , and according to the gray value change function of different horizontal columns Determine the corresponding horizontal column target gray value change rate function ; Extract the grayscale values ​​of all pixels in the corresponding column in the target stomach grayscale image in the longitudinal direction of the image with the longitudinal extraction step length i as the interval, and fit the grayscale values ​​of the corresponding row along the horizontal recognition direction of the image to form the longitudinal row target grayscale value change function of the corresponding row , and according to the gray value change function of different vertical rows Determine the corresponding vertical row target gray value change rate function ; Determine the gray value change rate function of the horizontal column target And the vertical row target gray value change rate function The horizontal target grayscale value change rate and the vertical target grayscale value change rate on the same pixel point are calculated, and the corresponding pixel point is marked as the target analysis pixel point; according to the gastric tumor image feature data, the pixel points on the target gastric grayscale image are marked in the following manner: for any target analysis pixel point, if there is a tumor boundary grayscale change rate range data in the gastric tumor image recognition feature data that satisfies the horizontal target grayscale value change rate of the target analysis pixel point belongs to the corresponding horizontal column tumor grayscale boundary change rate value range, and the vertical target grayscale value change rate of the target analysis pixel point belongs to the corresponding vertical row tumor grayscale boundary change rate value range. range, the target analysis pixel point is calibrated as a suspected tumor boundary pixel point; the suspected tumor boundary pixel points that are adjacent in the horizontal recognition direction and the vertical recognition direction of the image are connected, and the following gastric tumor analysis is performed: if the adjacent suspected tumor boundary pixel points are connected to form a closed curve, the closed curve is calibrated for the tumor position; if the adjacent suspected tumor boundary pixel points are connected and form a closed curve with the image boundary, the closed area is calibrated for the tumor position; if the adjacent suspected tumor boundary pixel points are connected but do not form a closed curve, and cannot form a closed curve with the image boundary, the target gastric image is determined to be a normal image.

[0053] When extracting image feature information obtained from big data, the target stomach image is first grayed, and the gray value change rate function is extracted on the grayed image in two directions at the same interval as the feature information extraction of the normal image and the tumor image. Then, the gray value change rate in the transverse direction and the gray value change rate in the longitudinal direction of the intersecting pixel points are compared and analyzed with the gastric tumor image feature data. The gray value change rate values ​​in the two directions are correlated. It is necessary to ensure that the gray value change rate in the two directions of the intersecting pixel points belongs to the range defined by the combination of the change rate range in the same transverse and longitudinal directions in the gastric tumor image feature data before the pixel point can be identified as a tumor boundary point. Of course, the determined pixel points are discrete points, while the tumor boundary is continuous, so after determining all the tumor boundary pixel points, it is necessary to determine whether these pixel points can form a reasonable boundary. It can be understood that if the pixel points formed by the intersection in the transverse and longitudinal directions are the boundaries of the tumor, then the pixel points are adjacent and can be connected. Considering the size of the extracted image area, it may not be able to completely cover the entire tumor area, so two situations are considered when judging. One is that the connection between adjacent pixels can form a closed curve, and the closed curve formed is the tumor area. The other is that the adjacent pixels can only form a curve, but this curve intersects with the boundary of the image and can form a closed curve with the boundary of the image. The area within the closed curve is the part of the tumor that appears in the image. It should be noted that the adjacency of pixel points needs to meet the adjacency in both the horizontal and vertical directions. Discontinuities in any direction cannot become the connection of a closed curve.

[0054] The present invention also provides a gastric tumor cell image extraction system, which is configured to: collect normal historical gastric image data, perform image feature extraction of the normal gastric state, and form normal gastric image feature data; collect historical gastric tumor image data, and perform gastric tumor image feature extraction in combination with the normal gastric image feature data, and form gastric tumor image feature data; obtain a target gastric image, and perform tumor recognition analysis in combination with the normal gastric image feature data and the gastric tumor image feature data, and form target image tumor feature information.

[0055] The system is configured to extract feature information from normal stomach image data and stomach tumor image data to provide identification of tumor feature information of target stomach images, thereby forming an efficient and accurate method for extracting feature information from stomach tumor images, laying an important material foundation for the extraction of feature information from stomach tumors, and ensuring the smooth and efficient extraction of feature information from stomach tumor images.

[0056] In summary, the gastric tumor cell image extraction method and system provided by the embodiments of the present invention have the following beneficial effects:

[0057] This method extracts the image feature information of the stomach under normal conditions by acquiring normal stomach image data, and at the same time acquires stomach tumor image data and extracts image feature information for the tumor. On this basis, the image feature information under normal conditions and tumor image feature information are used to further optimize the tumor image feature information, ensuring the rationality and accuracy of the tumor feature information, making the recognition of tumor features on the target stomach image more accurate and efficient. On the one hand, comprehensive and accurate tumor feature information can be established through big data analysis. On the other hand, compared with manual extraction of tumor feature information, identification with the help of tumor feature information obtained by big data is more efficient and accurate, providing important and accurate data reference for medical personnel's diagnosis and treatment.

[0058] The system is configured to extract feature information from normal stomach image data and stomach tumor image data to provide identification of tumor feature information of target stomach images, thereby forming an efficient and accurate method for extracting feature information from stomach tumor images, laying an important material foundation for the extraction of feature information from stomach tumors, and ensuring the smooth and efficient extraction of feature information from stomach tumor images.

[0059] In the embodiments of the present application, "indication" may include direct indication and indirect indication, and may also include explicit indication and implicit indication. The information indicated by a certain information is called information to be indicated. In the specific implementation process, there are many ways to indicate the information to be indicated, such as but not limited to, the information to be indicated can be directly indicated, such as the information to be indicated itself or the index of the information to be indicated. The information to be indicated can also be indirectly indicated by indicating other information, wherein there is an association between the other information and the information to be indicated. It is also possible to indicate only a part of the information to be indicated, while the other parts of the information to be indicated are known or agreed in advance. For example, the indication of specific information can also be achieved by means of the arrangement order of each piece of information agreed in advance (for example, specified by the protocol), thereby reducing the indication overhead to a certain extent. At the same time, the common parts of each piece of information can also be identified and indicated uniformly to reduce the indication overhead caused by indicating the same information separately.

[0060] In addition, the specific indication method may also be various existing indication methods, such as but not limited to the above-mentioned indication methods and various combinations thereof. The specific details of the various indication methods can refer to the prior art and will not be repeated herein. As can be seen from the above, for example, when it is necessary to indicate multiple information of the same type, different indication methods may be used for different information. In the specific implementation process, the desired indication method can be selected according to specific needs. The embodiment of the present application does not limit the selected indication method. In this way, the indication method involved in the embodiment of the present application should be understood to cover various methods that can enable the party to be indicated to obtain the information to be indicated.

[0061] It should be understood that the information to be indicated can be sent as a whole, or divided into multiple sub-information and sent separately, and the sending period and / or sending time of these sub-information can be the same or different. The specific sending method is not limited in the embodiment of the present application. Among them, the sending period and / or sending time of these sub-information can be pre-defined, for example, pre-defined according to a protocol, or can be configured by the sending end device by sending configuration information to the receiving end device.

[0062] "Pre-definition" or "pre-configuration" can be implemented by pre-saving corresponding codes, tables or other methods that can be used to indicate relevant information in the device, and the embodiments of the present application do not limit the specific implementation method. Among them, "saving" can mean saving in one or more memories. The one or more memories can be set separately or integrated in an encoder or decoder, a processor, or a communication device. The one or more memories can also be partially set separately and partially integrated in a decoder, a processor, or a communication device. The type of memory can be any form of storage medium, which is not limited by the embodiments of the present application.

[0063] The "protocol" involved in the embodiments of the present application may refer to a protocol family in the communication field, a standard protocol with a similar protocol family frame structure, or a related protocol used in future communication systems, and the embodiments of the present application do not make specific limitations on this.

[0064] In the embodiments of the present application, descriptions such as "when...", "in the case of...", "if" and "if" all mean that the device will make corresponding processing under certain objective circumstances. It does not limit the time, nor does it require the device to have a judgment action when implementing it, nor does it mean that there are other limitations.

[0065] In the description of the embodiments of the present application, unless otherwise specified, " / " indicates that the objects associated before and after are in an "or" relationship. For example, A / B can represent A or B; "and / or" in the embodiments of the present application is only a description of the association relationship of the associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. These three situations, where A and B can be singular or plural. In addition, in the description of the embodiments of the present application, unless otherwise specified, "multiple" refers to two or more than two. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, c can be single or multiple. In addition, in order to facilitate the clear description of the technical solution of the embodiments of the present application, in the embodiments of the present application, the words "first" and "second" are used to distinguish the same or similar items with basically the same functions and effects. Those skilled in the art will appreciate that words such as "first" and "second" do not limit the quantity and execution order, and words such as "first" and "second" do not necessarily limit the difference. At the same time, in the embodiments of the present application, words such as "exemplary" or "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "exemplary" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary" or "for example" is intended to present related concepts in a concrete way for easy understanding.

[0066] It should be understood that the processor in the embodiments of the present application may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0067] It should also be understood that the memory in the embodiments of the present application may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic random access memory (DRAM), synchronous DRAM (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0068] The above embodiments can be implemented in whole or in part by software, hardware (such as circuits), firmware or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center by wired (such as infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (for example, a floppy disk, a hard disk, a tape), an optical medium (for example, a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0069] It should be understood that the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. A and B can be singular or plural. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship, but it may also indicate an "and / or" relationship. Please refer to the context for specific understanding.

[0070] In this application, "at least one" means one or more, and "plurality" means two or more. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b, or c can mean: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple.

[0071] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0072] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0073] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0074] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0075] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0076] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0077] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, and other media that can store program codes.

[0078] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for extracting gastric tumor cell images, characterized in that: include: Collecting normal historical image data of the stomach, extracting image features of the normal state of the stomach, and forming normal image feature data of the stomach; Collecting gastric tumor historical image data, and extracting gastric tumor image features in combination with the gastric normal image feature data to form gastric tumor image feature data; Acquire a target stomach image, and perform tumor recognition analysis based on the normal stomach image feature data and the stomach tumor image feature data to form target image tumor feature information; The normal historical image data of the stomach is collected, and the image features of the normal state of the stomach are extracted to form the normal image feature data of the stomach, including: Performing grayscale processing on different normal historical stomach images in the normal historical stomach image data to form corresponding normal historical stomach grayscale images; Setting mutually orthogonal image transverse identification directions and image longitudinal identification directions, and extracting grayscale feature information in two directions on different normal historical images of the stomach, to form corresponding normal transverse grayscale feature information of the image and normal longitudinal grayscale feature information of the image; Gather the normal transverse grayscale feature information of the image and the normal longitudinal grayscale feature information of the image corresponding to different normal historical grayscale images of the stomach, perform normal feature analysis, and form the normal stomach image feature data; Setting mutually orthogonal image transverse identification directions and image longitudinal identification directions, and extracting grayscale feature information in two directions on different normal historical images of the stomach, respectively forming corresponding normal transverse grayscale feature information of the image and normal longitudinal grayscale feature information of the image, including: Set the horizontal extraction step length m, and extract the gray values ​​of all pixels in the corresponding column in the horizontal recognition direction of the image with the horizontal extraction step length m as an interval for different normal historical gray images of the stomach, and fit the gray values ​​of the corresponding column along the vertical recognition direction of the image to form a horizontal column normal gray value change function of the corresponding column. , wherein p represents the number of different normal historical grayscale images of the stomach, n represents the number of different columns sequentially extracted in the horizontal recognition direction of the image under the normal historical grayscale image of the stomach numbered p, and a is a natural number and represents the pixel number in the horizontal column; Set the extraction step length i in the longitudinal direction, and extract the grayscale values ​​of all pixels in the corresponding row in the longitudinal recognition direction of the image with the extraction step length i in the longitudinal direction as an interval for different normal grayscale images of the stomach, and fit the grayscale values ​​of the corresponding row along the transverse recognition direction of the image to form a normal grayscale value change function of the longitudinal row of the corresponding row. , wherein k represents the numbers of different rows sequentially extracted in the longitudinal recognition direction of the normal historical grayscale image of the stomach numbered p, and b is a natural number and represents the pixel number in the longitudinal row.

2. The method for extracting gastric tumor cell images according to claim 1, characterized in that: The normal transverse grayscale feature information of the image and the normal longitudinal grayscale feature information of the image corresponding to the different normal historical grayscale images of the stomach are collected to perform normal feature analysis to form the normal stomach image feature data, including: Determine the normal grayscale value change function of all the transverse columns in the normal historical grayscale images of the stomach The normal gray value change rate function of the horizontal column ,in, for The first derivative of ; For the normal gray value change rate function of different horizontal columns with the same arrangement number in different normal historical gray images of the stomach Delimit the range at each pixel point to form the normal grayscale change rate range data for each horizontal column ; Determine the normal grayscale value change function of all the vertical rows in the normal historical grayscale images of the stomach The vertical row normal gray value change rate function ,in, for The first derivative of ; For the normal grayscale value change rate function of different longitudinal rows with the same arrangement number in different normal historical grayscale images of the stomach The range is defined at each pixel point to form the normal gray value change rate range data for each vertical row .

3. The method for extracting gastric tumor cell images according to claim 2, characterized in that: The collecting of gastric tumor historical image data and extracting gastric tumor image features in combination with the gastric normal image feature data to form gastric tumor image feature data include: Performing grayscale processing on different gastric tumor historical images in the gastric tumor historical image data to form corresponding gastric tumor historical grayscale images; Extracting grayscale feature information in the horizontal identification direction and the vertical identification direction of the image from different historical grayscale images of the gastric tumor, and forming horizontal grayscale feature information of the image tumor and vertical grayscale feature information of the image tumor, respectively; Gathering the image tumor transverse grayscale feature information and the image tumor longitudinal grayscale feature information corresponding to different gastric tumor historical grayscale images, performing tumor feature analysis, and forming preliminary gastric tumor image feature data; The gastric tumor recognition feature data is extracted by combining the normal stomach image feature data and the preliminary stomach tumor image feature data to form the stomach tumor image feature data.

4. The method for extracting gastric tumor cell images according to claim 3, characterized in that: The step of extracting grayscale feature information of different gastric tumor historical grayscale images in the horizontal identification direction and the vertical identification direction of the image to form horizontal grayscale feature information of the image tumor and vertical grayscale feature information of the image tumor respectively includes: For different gastric tumor historical grayscale images, the grayscale values ​​of the corresponding columns are extracted in sequence with the horizontal direction extraction step length m as an interval in the horizontal recognition direction of the image, and the grayscale values ​​of the corresponding columns along the vertical recognition direction of the image are fitted to form a horizontal column tumor grayscale value change function of the corresponding column. , wherein q represents the numbers of different gastric tumor historical grayscale images, u represents the numbers of different columns sequentially extracted in the horizontal recognition direction of the image under the gastric tumor historical grayscale image numbered q, and u=n; For different gastric tumor historical grayscale images, the grayscale values ​​of all pixels in the corresponding row are extracted in sequence in the longitudinal recognition direction of the image with the longitudinal extraction step length i as the interval, and the grayscale values ​​of the corresponding row along the transverse recognition direction of the image are fitted to form a longitudinal tumor grayscale value change function of the corresponding row. , wherein v represents the numbers of different rows sequentially extracted in the longitudinal identification direction of the image under the historical grayscale image of the gastric tumor numbered q, and v=k.

5. The method for extracting gastric tumor cell images according to claim 4, characterized in that: The collecting of the transverse grayscale feature information of the image tumor and the longitudinal grayscale feature information of the image tumor corresponding to different historical grayscale images of the gastric tumor, performing tumor feature analysis, and forming preliminary feature data of the gastric tumor image includes: Determine the grayscale value change function of all the transverse columns of the tumor in the different gastric tumor historical grayscale images The gray value change rate function of the horizontal column tumor ,in, for The first derivative of ; For different said transverse columns of tumors with the same arrangement number in different said gastric tumor historical grayscale images, the grayscale value change rate function The range is defined at each pixel point to form the range data of tumor grayscale change rate corresponding to each horizontal column. ; Determine the grayscale value change function of all the longitudinal rows of tumors in different gastric tumor historical grayscale images The function of the change rate of the longitudinal tumor gray value ,in, for The first derivative of ; The gray value change rate function of different longitudinal rows of tumors with the same arrangement number in different gastric tumor historical gray images is The range is delineated at each pixel point to form the range data of the gray value change rate of the longitudinal tumor corresponding to each longitudinal row. .

6. The method for extracting gastric tumor cell images according to claim 5, characterized in that: The step of combining the normal stomach image feature data and the preliminary stomach tumor image feature data to extract stomach tumor recognition features and form the stomach tumor image feature data includes: For the normal grayscale change rate range data of the horizontal columns with the same arrangement number And the horizontal column tumor grayscale change rate range data , compare the change rate range on each pixel point a in turn, and determine the tumor grayscale change rate range data of the horizontal column at the pixel point The horizontal column tumor gray value range and the horizontal column normal gray change rate range data The horizontal column normal grayscale value range does not intersect the horizontal column tumor grayscale value range, marking the corresponding pixel point as a horizontal suspected tumor boundary point, and marking the corresponding horizontal column tumor grayscale value range as a horizontal column suspected tumor grayscale boundary change rate value range; For the normal gray value change rate range data of the vertical rows with the same arrangement number and the longitudinal row tumor gray value change rate range data , compare the change rate range of each pixel point b in turn, and determine the change rate range data of the longitudinal tumor gray value at the pixel point The vertical row tumor gray value range and the vertical row normal gray value change rate range data The vertical row of normal grayscale value ranges does not intersect the vertical row of tumor grayscale value ranges, marking the corresponding pixel points as vertical suspected tumor boundary points, and marking the corresponding vertical row of tumor grayscale value ranges as vertical row suspected tumor grayscale boundary change rate value ranges; Determine the pixel position where the transverse suspected tumor boundary point and the longitudinal suspected tumor boundary point are the same pixel point, and mark it as the tumor boundary pixel position; Extracting the grayscale boundary change rate value range of the suspected tumor in the transverse column corresponding to the tumor boundary pixel position, and marking it as the grayscale boundary change rate value range of the tumor in the transverse column; Extracting the grayscale boundary change rate value range of the longitudinal row of suspected tumors corresponding to the tumor boundary pixel position, and marking it as the grayscale boundary change rate value range of the longitudinal row of tumors; Combining the horizontal column tumor grayscale boundary change rate value range and the vertical row tumor grayscale boundary change rate value range to form tumor boundary grayscale change rate range data; All the grayscale change rate range data of the tumor boundary are collected to form the gastric tumor image feature data.

7. The method for extracting gastric tumor cell images according to claim 6, characterized in that: The step of acquiring the target stomach image and combining the normal stomach image feature data and the stomach tumor image feature data to perform tumor recognition analysis to form target image tumor feature information includes: Performing grayscale processing on the target stomach image to form a target stomach grayscale image; The target stomach grayscale image is extracted in the horizontal direction of the image with the horizontal extraction step length m as an interval, and the grayscale values ​​of all pixels in the corresponding column are extracted in sequence, and the grayscale values ​​of the corresponding column along the vertical recognition direction of the image are fitted to form the horizontal column target grayscale value change function of the corresponding column. , and according to the different horizontal column target gray value change functions Determine the corresponding horizontal column target gray value change rate function ; The target stomach grayscale image is extracted in the longitudinal direction of the image with the longitudinal extraction step i as the interval, the grayscale values ​​of all pixels in the corresponding column are extracted in sequence, and the grayscale values ​​of the corresponding row along the transverse recognition direction of the image are fitted to form the longitudinal row target grayscale value change function of the corresponding row , and according to the gray value change function of different vertical rows of targets Determine the corresponding vertical row target gray value change rate function ; Determine the horizontal column target gray value change rate function and the vertical row target gray value change rate function The horizontal target grayscale value change rate and the vertical target grayscale value change rate located at the same pixel point on the image processing unit are calculated, and the corresponding pixel point is marked as the target analysis pixel point; According to the gastric tumor image feature data, the pixel points on the target gastric grayscale image are marked in the following manner: For any of the target analysis pixels, if there is the tumor boundary grayscale change rate range data in the gastric tumor image recognition feature data that satisfies the fact that the horizontal target grayscale value change rate of the target analysis pixel belongs to the corresponding horizontal column tumor grayscale boundary change rate value range, and the vertical target grayscale value change rate of the target analysis pixel belongs to the corresponding vertical row tumor grayscale boundary change rate value range, then the target analysis pixel is marked as a suspected tumor boundary pixel; Connect the suspected tumor boundary pixel points that are adjacent in both the horizontal identification direction of the image and the vertical identification direction of the image, and perform the following gastric tumor analysis: If the adjacent suspected tumor boundary pixel points are connected to form a closed curve, the closed curve is calibrated for the tumor position; If the adjacent suspected tumor boundary pixel points are connected and form a closed curve with the image boundary, the closed area is calibrated for the tumor position; If the adjacent suspected tumor boundary pixel points are connected to each other but fail to form a closed curve, and cannot form a closed curve with the image boundary, then the target stomach image is determined to be a normal image.

8. A gastric tumor cell image extraction system, characterized in that: Configured as: Collecting normal historical image data of the stomach, extracting image features of the normal state of the stomach, and forming normal image feature data of the stomach; Collecting gastric tumor historical image data, and extracting gastric tumor image features in combination with the gastric normal image feature data to form gastric tumor image feature data; Acquire a target stomach image, and perform tumor recognition analysis based on the normal stomach image feature data and the stomach tumor image feature data to form target image tumor feature information; The normal historical image data of the stomach is collected, and the image features of the normal state of the stomach are extracted to form the normal image feature data of the stomach, including: Performing grayscale processing on different normal historical stomach images in the normal historical stomach image data to form corresponding normal historical stomach grayscale images; Setting mutually orthogonal image transverse identification directions and image longitudinal identification directions, and extracting grayscale feature information in two directions on different normal historical images of the stomach, to form corresponding normal transverse grayscale feature information of the image and normal longitudinal grayscale feature information of the image; Gather the normal transverse grayscale feature information of the image and the normal longitudinal grayscale feature information of the image corresponding to different normal historical grayscale images of the stomach, perform normal feature analysis, and form the normal stomach image feature data; Setting mutually orthogonal image transverse identification directions and image longitudinal identification directions, and extracting grayscale feature information in two directions on different normal historical images of the stomach, respectively forming corresponding normal transverse grayscale feature information of the image and normal longitudinal grayscale feature information of the image, including: Set the horizontal extraction step length m, and extract the gray values ​​of all pixels in the corresponding column in the horizontal recognition direction of the image with the horizontal extraction step length m as an interval for different normal historical gray images of the stomach, and fit the gray values ​​of the corresponding column along the vertical recognition direction of the image to form a horizontal column normal gray value change function of the corresponding column. , wherein p represents the number of different normal historical grayscale images of the stomach, n represents the number of different columns sequentially extracted in the horizontal recognition direction of the image under the normal historical grayscale image of the stomach numbered p, and a is a natural number and represents the pixel number in the horizontal column; Set the extraction step length i in the longitudinal direction, and extract the grayscale values ​​of all pixels in the corresponding row in the longitudinal recognition direction of the image with the extraction step length i in the longitudinal direction as an interval for different normal grayscale images of the stomach, and fit the grayscale values ​​of the corresponding row along the transverse recognition direction of the image to form a normal grayscale value change function of the longitudinal row of the corresponding row. , wherein k represents the numbers of different rows sequentially extracted in the longitudinal recognition direction of the normal historical grayscale image of the stomach numbered p, and b is a natural number and represents the pixel number in the longitudinal row.

Images