Multi-modal fusion maternal and child medical image intelligent identification and diagnosis auxiliary system and method thereof

Abstract

The invention discloses a multi-modal fusion maternal and child medical image intelligent identification and diagnosis auxiliary system and method, and belongs to the technical field of medical health, and the system comprises a data obtaining module which is used for obtaining a medical image of a related part of a patient, the related part comprises an area corresponding to gland tissue, related processing is carried out on the medical image, and the medical image is obtained; the correlation processing includes distinguishing the medical image into an image corresponding to a gland and an image corresponding to connective tissue. According to the method, by obtaining and distinguishing gland and connective tissue medical images, extracting edge features and monitoring area changes, the lesion site can be accurately judged, the distance change between the blood vessel and the blood vessel group can be quantitatively evaluated, then the lesion development trend is scientifically judged, a comprehensive and objective diagnosis basis is provided for doctors, and the diagnosis efficiency is improved. And the diagnosis accuracy and the clinical decision scientificity are effectively improved.

Description

Technical Field

[0001] This invention relates to the field of medical and health technology, and in particular to a multimodal fusion intelligent recognition and diagnostic assistance system and method for maternal and child medical images. Background Technology

[0002] In real life, according to medical statistics, the most common malignant diseases among women are, in descending order, breast cancer, lung cancer, colorectal cancer, stomach cancer, and thyroid cancer. Modern medical technology allows for the diagnosis and treatment of these diseases through medical imaging.

[0003] Regarding this research, application CN202110165179.6 provides a multimodal medical image intelligent assisted diagnosis and treatment system. This technical solution includes an acquisition unit and a diagnosis and treatment information management unit. The diagnosis and treatment information management unit includes an analysis unit and a diagnosis and treatment unit. The acquisition unit collects multiple comparative cases, and after image processing and category analysis by the processing and integration unit, the data is sent to the analysis unit. After the optimization unit analyzes the biological characteristics of the comparative cases, the analysis unit uses an algorithm unit to perform AI prognostic algorithm detection on the comparative cases. This technical solution solves the problems of low efficiency in hospital image reading and the inability to predict early-stage cancer.

[0004] Another application, CN202411592524.4, discloses a fetal assisted diagnostic system based on multimodal medical information fusion. This technical solution includes a signal preprocessing module, an image generation module, a text extraction module, a multimodal fusion module, and an output module. The multimodal fusion module utilizes an image-text feature fusion network (ITFN) to extract features from image data acquired by the image generation module and text data acquired by the text extraction module, obtaining image feature vectors Mi and Mt. The image feature vectors Mi and Mt are then weighted and fused to obtain a multimodal feature vector Zf. The output module uses a fully connected layer to reduce the dimensionality of the multimodal feature vector Zf and outputs the classification results for normal and pathological samples. This technical solution can effectively reduce the misdiagnosis rate of fetal distress pathology and has certain practical application value.

[0005] However, the above-mentioned technical solutions mostly adopt static fusion strategies, which make it difficult to dynamically adjust the modal contribution based on image quality or diagnostic tasks (such as breast BI-RADS grading). This may result in the masking of key structural features (such as calcification features), making it difficult to meet the needs of dynamic monitoring. Summary of the Invention

[0006] In view of the problems existing in the field of medical and health technology, the present invention is proposed.

[0007] Therefore, one of the objectives of this invention is to provide a multimodal fusion intelligent recognition and diagnostic assistance system and method for maternal and child medical images. By acquiring and distinguishing medical images of glands and connective tissues, extracting edge features and monitoring area changes, it can accurately determine the lesion site and quantitatively assess the changes in the distance between blood vessels and blood vessel groups, thereby scientifically judging the development trend of the lesion, providing doctors with comprehensive and objective diagnostic basis, and effectively improving the accuracy of diagnosis and the scientific nature of clinical decision-making.

[0008] To solve the above-mentioned technical problems, the present invention provides the following technical solution: On the one hand, this invention provides a multimodal fusion-based intelligent recognition and diagnostic assistance system for maternal and child medical images, comprising: The data acquisition module is used to acquire medical images of relevant parts of the patient, including areas corresponding to glandular tissue, and to perform relevant processing on the medical images, including distinguishing the medical images into images corresponding to glands and images corresponding to connective tissue. The data processing module includes an acquisition unit and a feature extraction unit; The acquisition unit responds to the image corresponding to the gland and the image corresponding to the connective tissue, and is used to acquire the edges of the gland and the connective tissue in the image corresponding to the gland and the image corresponding to the connective tissue, and to determine whether the edges of the gland and the connective tissue are clear or blurry. The feature extraction unit is used to extract features from blurred edges in glands and connective tissue. The extraction steps include: Divide the edge of the gland and connective tissue into 6 to 10 edge points, and give a monitoring point based on each edge point. The monitoring point is the monitoring point that is closest to and clear to the corresponding edge point, and calculate the distance between the monitoring point and the corresponding edge point. Given the outline of each monitoring point, calculate the area of ​​each monitoring point; A preset diagnostic cycle is defined, wherein each diagnostic cycle consists of 7 to 10 days, and the change characteristics of the area are obtained in each diagnostic cycle. The data fusion and analysis module includes an acquisition unit, a comparison unit, and a judgment unit. The acquisition unit is used to acquire relevant information based on the change characteristics. The relevant information includes density, which is the density of the monitoring points corresponding to the gland and the connective tissue when the area of ​​the monitoring points corresponding to each edge point of the gland and each edge point of the connective tissue is changing in a shrinking trend. The comparison unit responds to the acquisition unit and is used to compare the density. The determination unit is used to make a determination based on the comparison results.

[0009] In a preferred embodiment of the present invention, in the feature extraction unit, the contour of each monitoring point is given, and the area of ​​each monitoring point is calculated according to the following formula: ; In the formula, Represents the area of ​​the polygon; Indicates the number of vertices of the polygon; Represents the polygon's first... The coordinates of the vertices, where, =1, 2, ..., .

[0010] In a preferred embodiment of the present invention, in the determination unit, if the density of monitoring points corresponding to the gland is greater than the density of monitoring points corresponding to connective tissue, the system determines that the patient's lesion site is concentrated in the gland; otherwise, no determination is made.

[0011] In a preferred embodiment of the present invention: if the system determines that the patient's lesion is concentrated in the gland, relevant monitoring is performed at the monitoring point corresponding to the gland. The relevant monitoring includes monitoring the distribution of blood vessels connected to the monitoring point, collecting blood vessels with disordered distribution in the distribution, including blood vessels with cross distribution, marking the blood vessels with cross distribution as blood vessel groups, counting the number of blood vessel groups, and obtaining the cross point in each blood vessel group, the cross point being the cross point at the center of the blood vessel group; based on the cross point, collecting the blood vessel closest to the corresponding blood vessel group, and obtaining the regular change of the distance between the blood vessel and the corresponding blood vessel group based on 3 consecutive diagnostic cycles.

[0012] In a preferred embodiment of the present invention, the regular changes of the blood vessels and blood vessel groups are obtained and calculated according to the following formula: ; In the formula, Indicates the first The vascular group in the first The diagnostic cycle and the first - The change in distance over one diagnostic cycle; Indicates the first In the first diagnostic cycle The distance from the center of a vascular group to a point on the nearest blood vessel corresponding to that vascular group; Indicates the first -1 diagnostic cycle The distance from the center of a vascular group to a point on the nearest blood vessel corresponding to that vascular group.

[0013] In a preferred embodiment of the present invention, the following formula is also included: ; In the formula, express The average distance change rate of each vascular group in two adjacent diagnostic cycles; This indicates the total number of blood vessels in the group; Indicates the first The vascular group in the first The diagnostic cycle and the first - The change in distance over one diagnostic cycle; Indicates the first In the first diagnostic cycle The distance from the center of a vascular group to a point on the nearest blood vessel corresponding to that vascular group; Indicates the first -1 diagnostic cycle The distance from the center of a vascular group to a point on the nearest blood vessel corresponding to that vascular group.

[0014] In a preferred embodiment of the present invention, if the distance between the blood vessel and the corresponding blood vessel group decreases according to the calculation results, the system determines that the lesion at the gland is developing towards a worsening trend; otherwise, no determination is made.

[0015] In a preferred embodiment of the present invention, if the system determines that the lesion at the gland is developing towards a worsening trend, then while obtaining the regular changes in the distance between the blood vessel and the corresponding blood vessel group based on three consecutive diagnostic cycles, the system also obtains the changes in the blood vessel group. If the blood vessels in the blood vessel group show a trend of dispersion, then the system does not determine that the lesion at the gland is developing towards a worsening trend; otherwise, it does.

[0016] On the other hand, the present invention provides a method for applying the multimodal fusion-based intelligent recognition and diagnostic assistance system for maternal and child medical images as described above, the steps of which include: Acquire medical images of relevant parts of the patient, including areas corresponding to glandular tissue, and perform relevant processing on the medical images, including distinguishing the medical images into images corresponding to glands and images corresponding to connective tissue; The edges of glands and connective tissues are acquired from images corresponding to glands and images corresponding to connective tissues, and the clarity or blurriness of the edges of glands and connective tissues is determined. Feature extraction is performed on blurred edges in glands and connective tissue. The extraction steps include: Divide the edge of the gland and connective tissue into 6 to 10 edge points, and give a monitoring point based on each edge point. The monitoring point is the monitoring point that is closest to and clear to the corresponding edge point, and calculate the distance between the monitoring point and the corresponding edge point. Given the outline of each monitoring point, calculate the area of ​​each monitoring point; A preset diagnostic cycle is defined, wherein each diagnostic cycle consists of 7 to 10 days, and the change characteristics of the area are obtained in each diagnostic cycle. Based on the change characteristics, relevant information is obtained, including density, which is the degree of density of the monitoring points corresponding to the gland and the connective tissue when the area of ​​the monitoring points corresponding to each edge point of the gland and each edge point of the connective tissue is changing in a shrinking trend. The density levels are compared. The judgment is made based on the comparison results.

[0017] Beneficial effects: 1. By acquiring medical images of relevant parts of the patient, especially distinguishing between glandular and connective tissues, the system helps doctors to more accurately locate lesion areas. By analyzing the edge features of glands and connective tissues, especially the precise extraction of blurred edges and the monitoring of area change features, it provides more reliable reference information for diagnosis. 2. Through the collaborative work of the acquisition unit, comparison unit, and judgment unit, the system can quickly locate the lesion area based on the density changes of monitoring points in glands and connective tissues. When the system determines that the lesion is concentrated in the gland, it will further monitor the distribution of blood vessels connected to the gland monitoring points. In particular, for disordered blood vessel groups, by counting the number of blood vessel groups and obtaining the intersection points, the system can more comprehensively assess the severity of the lesion and the possible trend of deterioration. 3. By calculating the changes in distance between blood vessels and vascular groups during continuous diagnostic cycles, the system can quantify the development process of lesions, which helps improve the scientific rigor and objectivity of diagnosis and assists doctors in developing more reasonable treatment plans. Furthermore, by analyzing the changing patterns of distance between blood vessels and vascular groups, as well as the dispersion trend of vascular groups, the system can comprehensively judge the development trend of lesions in glands, which helps doctors intervene in advance to prevent the lesions from worsening, thereby improving the treatment effect and quality of life of patients. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein: Figure 1 This is a schematic diagram of the modular structure of the multimodal fusion intelligent recognition and diagnostic assistance system for maternal and child medical images according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the method flow according to an embodiment of the present invention; The diagram is labeled as follows: 110 - Data acquisition module; 120 - Data processing module; 1201 - Acquisition unit; 1202 - Feature extraction unit; 130 - Data fusion and analysis module; 1301 - Acquisition unit; 1302 - Comparison unit; 1303 - Judgment unit. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the described embodiments of the present invention are within the scope of protection of the present invention.

[0020] Because existing technical solutions mostly adopt static fusion strategies, they may obscure key structural features (such as calcification features), making it difficult to meet the needs of dynamic monitoring.

[0021] Based on this, the present invention proposes a multimodal fusion intelligent recognition and diagnostic assistance system and method for maternal and child medical images. By acquiring and distinguishing medical images of glands and connective tissues, extracting edge features and monitoring area changes, it can accurately determine the lesion site and quantitatively assess the changes in the distance between blood vessels and blood vessel groups, thereby scientifically judging the development trend of the lesion, providing doctors with comprehensive and objective diagnostic basis, and effectively improving the accuracy of diagnosis and the scientific nature of clinical decision-making.

[0022] The present solution will be further described in detail below through embodiments and in conjunction with the accompanying drawings.

[0023] Reference Figures 1 to 2 As one embodiment of the present invention, this embodiment provides a multimodal fusion-based intelligent recognition and diagnostic assistance system for maternal and child medical images, comprising: The data acquisition module 110 is used to acquire medical images of relevant parts of the patient, including areas corresponding to glandular tissue, and to perform relevant processing on the medical images, including distinguishing the medical images into images corresponding to glands and images corresponding to connective tissue. It is important to note that glandular tissue is a crucial site for observing whether a woman has breast cancer. When cells in glandular tissue proliferate abnormally or mutate, tumors may form, eventually developing into breast cancer. Therefore, the health of the glands is essential for assessing breast cancer risk. The connective tissue matrix surrounds the breast, providing support and fixation. Although connective tissue itself is not the direct site of breast cancer, abnormal proliferation or fibrosis of connective tissue is associated with much of the pathological process in breast cancer. Therefore, distinguishing between images corresponding to glandular tissue and images corresponding to connective tissue is of practical significance.

[0024] Meanwhile, by distinguishing between glandular and connective tissue images, a clear anatomical basis is provided for subsequent analysis, reducing the risk of misdiagnosis. This lays the foundation for subsequent multi-dimensional information such as fused images, edge features, and vascular distribution, thereby improving the comprehensiveness of diagnosis.

[0025] Data processing module 120, which includes acquisition unit 1201 and feature extraction unit 1202; The acquisition unit 1201 responds to the image corresponding to the gland and the image corresponding to the connective tissue, and is used to acquire the edges of the gland and the connective tissue in the image corresponding to the gland and the image corresponding to the connective tissue, and to determine whether the edges of the gland and the connective tissue are clear or blurry. It should be noted that the clarity or blurriness (blurriness) of images of glands and connective tissue obtained from patients indicates the presence or development of certain diseases. For example, some diseases may cause tissue edema, inflammation, or fibrosis, thus affecting the clarity of the images. Therefore, it is of practical significance to determine the clarity or blurriness of the edges of glands and connective tissue.

[0026] The determination of clarity and blurriness is based on whether the edges of the glands and connective tissues are clear. If the edges are clear, they are not considered blurry.

[0027] Feature extraction unit 1202 is used to extract features from blurred edges in glands and connective tissue. The extraction steps include: Divide the edge of the gland and connective tissue into 6 to 10 edge points, and give a monitoring point based on each edge point. The monitoring point is the monitoring point that is closest to and clear to the corresponding edge point, and calculate the distance between the monitoring point and the edge point. Given the outline of each monitoring point, calculate the area of ​​each monitoring point; The diagnostic cycle is preset, with each diagnostic cycle consisting of 7 to 10 days. The change characteristics of the area are obtained in each diagnostic cycle. In this embodiment, by monitoring changes in the distance and area of ​​monitoring points, minute lesions (such as glandular tissue atrophy or hyperplasia) are captured, thereby improving the sensitivity of early diagnosis. Periodically tracking changes in area provides data support for understanding the progression of lesions.

[0028] The data fusion analysis module 130 includes an acquisition unit 1301, a comparison unit 1302, and a judgment unit 1303. The acquisition unit 1301 is used to acquire relevant information based on the change characteristics. The relevant information includes density, which is the degree of density of the monitoring points corresponding to the edge points of the gland and the monitoring points corresponding to the edge points of the connective tissue when the area of ​​the monitoring points corresponding to the edge points of the gland and the connective tissue is changing in a shrinking trend. It should be noted that if the number of monitoring points corresponding to the edge points of connective tissue is greater than the number of monitoring points corresponding to the edge points of glands, it indicates that the density of monitoring points corresponding to connective tissue is greater than that corresponding to glands.

[0029] The comparison unit 1302 is a response acquisition unit used for comparing the density. The determination unit 1303 is used to make a determination based on the comparison results; By comparing density, the lesion area (such as abnormal breast tissue) can be quickly located, reducing missed diagnoses; This quantifies density indicators, reducing subjective judgment errors.

[0030] In the feature extraction unit, given the contour of each monitoring point, the area of ​​each monitoring point is calculated using the following formula: ; In the formula, Represents the area of ​​the polygon; Indicates the number of vertices of the polygon; Represents the polygon's first... The coordinates of the vertices, where, =1, 2, ..., .

[0031] In the judgment unit, if the density of monitoring points corresponding to glands is greater than that of monitoring points corresponding to connective tissue, the system determines that the patient's lesion site is concentrated in the glands; otherwise, no judgment is made.

[0032] If the system determines that the patient's lesion is concentrated in the gland, relevant monitoring is performed at the monitoring point corresponding to the gland. The relevant monitoring includes monitoring the distribution of blood vessels connected to the monitoring point, collecting blood vessels with disordered distribution, including those with intersecting distribution, marking the intersecting blood vessels as blood vessel groups, counting the number of blood vessel groups, and obtaining the intersection point in each blood vessel group. The intersection point is the intersection point at the center of the blood vessel group. Based on the intersection point, the blood vessel closest to the corresponding blood vessel group is collected, and the regular changes in the distance between the blood vessel and the corresponding blood vessel group are obtained based on three consecutive diagnostic cycles. In this embodiment, three or more intersecting blood vessels constitute a blood vessel group; Clinical experience shows that breast cancer cells release tumor angiogenesis factor (TAF), which stimulates capillary growth and creates a rich vascular network in the tumor area. These new blood vessels penetrate from the periphery of the tumor inwards, constantly renewing and redistributing as the tumor grows, resulting in an increase in the number of blood vessels and a disordered distribution.

[0033] Disordered angiogenesis (such as tumor neovascularization) is an important marker of lesions. The system can provide early warning of malignant trends and quantify the rate of lesion progression (such as the depth of vascular invasion) through periodic distance changes.

[0034] The regular changes in blood vessels and vascular groups are obtained and calculated using the following formula: ; In the formula, Indicates the first The vascular group in the first The diagnostic cycle and the first - The change in distance over one diagnostic cycle; Indicates the first In the first diagnostic cycle The distance from the center of a vascular group to a point on the nearest blood vessel corresponding to that vascular group; Indicates the first -1 diagnostic cycle The distance from the center of a vascular group to a point on the nearest blood vessel corresponding to that vascular group.

[0035] In addition to the above, it also includes calculations based on the following formula: ; In the formula, express The average distance change rate of each vascular group in two adjacent diagnostic cycles; This indicates the total number of blood vessels in the group; Indicates the first The vascular group in the first The diagnostic cycle and the first - The change in distance over one diagnostic cycle; Indicates the first In the first diagnostic cycle The distance from the center of a vascular group to a point on the nearest blood vessel corresponding to that vascular group; Indicates the first -1 diagnostic cycle The distance from the center of a vascular group to a point on the nearest blood vessel corresponding to that vascular group; It should be noted that, combining the two calculation formulas above, the change in distance can reflect the change in the distance between a single vascular group and its corresponding vessel, while the average rate of change in distance can comprehensively assess the overall trend of change in the distance between all vascular groups and their corresponding vessels. If The larger distance indicates that the distance between blood vessels and blood vessel groups is changing significantly.

[0036] This can objectively assess the rate of vascular invasion and help determine the degree of malignancy, while continuous periodic analysis can detect the accelerated (e.g., continuously decreasing distance) or stable (e.g., fluctuating distance) state of the lesion in advance.

[0037] Furthermore, based on the calculation results, if the distance between the blood vessel and the corresponding blood vessel group decreases, the system determines that the lesion at the gland is developing towards a worsening trend; otherwise, no determination is made.

[0038] Meanwhile, if the system determines that the lesion at the gland is developing towards a worsening trend, it will acquire the changes in the vascular group while acquiring the regular changes in the distance between the blood vessel and the corresponding vascular group based on three consecutive diagnostic cycles. If the blood vessels in the vascular group show a trend of dispersion, the system will not determine that the lesion at the gland is developing towards a worsening trend; otherwise, it will.

[0039] If the distance between blood vessels and vascular groups shows a decreasing trend, and the vascular groups do not disperse (increased vascular density), then the lesion is judged to be worsening. This combination of distance changes and vascular group morphology (dispersion / density) avoids false positives caused by abnormalities in a single indicator, and at the same time provides doctors with a clear basis for whether intervention (such as biopsy or surgery) is needed, thus optimizing the treatment plan.

[0040] Based on the above, this application achieves comprehensive diagnosis from anatomical structure to functional changes through a multimodal fusion process of image segmentation, edge feature extraction, vascular dynamic monitoring, and quantitative trend analysis. This can improve the sensitivity of early diagnosis (detection of minute lesions) and quantify the speed of lesion progression, while reducing subjective errors.

[0041] This embodiment, in conjunction with the aforementioned multimodal fusion-based intelligent recognition and diagnostic assistance system for maternal and child medical images, also proposes a working method for this system, as follows: S10: Acquire medical images of relevant parts of the patient, including areas corresponding to glandular tissue, and perform relevant processing on the medical images, including distinguishing the medical images into images corresponding to glands and images corresponding to connective tissue. S20: Collect the edges of glands and connective tissue in the images corresponding to glands and connective tissue, and determine whether the edges of glands and connective tissue are clear or blurred. S30: Feature extraction of blurred edges in glands and connective tissue. The extraction steps include: S301: Divide the edge of the gland and connective tissue into 6 to 10 edge points, give a monitoring point based on each edge point, the monitoring point is the monitoring point that is closest to and clear to the corresponding edge point, and calculate the distance between the monitoring point and the corresponding edge point; S302: Given the outline of each monitoring point, calculate the area of ​​each monitoring point; S303: Preset diagnostic cycle, the diagnostic cycle includes a diagnostic cycle of 7 to 10 days, and obtain the area change characteristics in each diagnostic cycle; S40: Obtain relevant information based on the change characteristics. The relevant information includes density, which is the degree of density of monitoring points corresponding to each edge point of the gland and each edge point of the connective tissue when the area of ​​the monitoring points corresponding to each edge point of the gland and the connective tissue is changing in a shrinking trend. S50: Comparison of density levels; S60: Make a judgment based on the comparison results.

[0042] In summary, by acquiring and distinguishing medical images of glands and connective tissues, extracting edge features, and monitoring area changes, this invention can accurately determine the location of lesions and quantitatively assess changes in the distance between blood vessels and vascular groups, thereby scientifically judging the development trend of lesions and providing doctors with comprehensive and objective diagnostic evidence, effectively improving diagnostic accuracy and the scientific nature of clinical decision-making.

[0043] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A multimodal fusion-based intelligent recognition and diagnostic assistance system for maternal and child medical images, characterized in that, include: The data acquisition module is used to acquire medical images of relevant parts of the patient, including areas corresponding to glandular tissue, and to perform relevant processing on the medical images, including distinguishing the medical images into images corresponding to glands and images corresponding to connective tissue. The data processing module includes an acquisition unit and a feature extraction unit; The acquisition unit responds to the image corresponding to the gland and the image corresponding to the connective tissue, and is used to acquire the edges of the gland and the connective tissue in the image corresponding to the gland and the image corresponding to the connective tissue, and to determine whether the edges of the gland and the connective tissue are clear or blurry. The feature extraction unit is used to extract features from blurred edges in glands and connective tissue. The extraction steps include: Divide the edge of the gland and connective tissue into 6 to 10 edge points, and give a monitoring point based on each edge point. The monitoring point is the monitoring point that is closest to and clear to the corresponding edge point, and calculate the distance between the monitoring point and the corresponding edge point. Given the outline of each monitoring point, calculate the area of ​​each monitoring point; A preset diagnostic cycle is defined, wherein each diagnostic cycle consists of 7 to 10 days, and the change characteristics of the area are obtained in each diagnostic cycle. The data fusion and analysis module includes an acquisition unit, a comparison unit, and a judgment unit. The acquisition unit is used to acquire relevant information based on the change characteristics. The relevant information includes density, which is the density of the monitoring points corresponding to the gland and the connective tissue when the area of ​​the monitoring points corresponding to each edge point of the gland and each edge point of the connective tissue is changing in a shrinking trend. The comparison unit responds to the acquisition unit and is used to compare the density. The determination unit is used to make a determination based on the comparison results.

2. The multimodal fusion intelligent recognition and diagnostic assistance system for maternal and child medical images as described in claim 1, characterized in that, In the feature extraction unit, given the contour of each monitoring point, the area of ​​each monitoring point is calculated according to the following formula: ； In the formula, Represents the area of ​​the polygon; Indicates the number of vertices of the polygon; Represents the polygon's first... The coordinates of the vertices, where, =1, 2, ..., .

3. The multimodal fusion intelligent recognition and diagnostic assistance system for maternal and child medical images as described in claim 1, characterized in that, In the determination unit, if the density of monitoring points corresponding to the gland is greater than the density of monitoring points corresponding to connective tissue, the system determines that the patient's lesion is concentrated in the gland; otherwise, no determination is made.

4. The multimodal fusion intelligent recognition and diagnostic assistance system for maternal and child medical images as described in claim 3, characterized in that, If the system determines that the patient's lesion is concentrated in the gland, relevant monitoring is performed at the monitoring point corresponding to the gland. The relevant monitoring includes monitoring the distribution of blood vessels connected to the monitoring point, collecting blood vessels with disordered distribution in the distribution, including blood vessels with intersecting distribution, marking the intersecting blood vessels as blood vessel groups, counting the number of blood vessel groups, and obtaining the intersection point in each blood vessel group, the intersection point being the intersection point at the center of the blood vessel group; based on the intersection point, collecting the blood vessel closest to the corresponding blood vessel group, and obtaining the regular change of the distance between the blood vessel and the corresponding blood vessel group based on 3 consecutive diagnostic cycles.

5. The multimodal fusion intelligent recognition and diagnostic assistance system for maternal and child medical images as described in claim 4, characterized in that, The regular changes of the blood vessels and blood vessel groups were obtained and calculated according to the following formula: ； In the formula, Indicates the first The vascular group in the first The diagnostic cycle and the first - The change in distance over one diagnostic cycle; Indicates the first In the first diagnostic cycle The distance from the center of a vascular group to a point on the nearest blood vessel corresponding to that vascular group; Indicates the first -1 diagnostic cycle The distance from the center of a vascular group to a point on the nearest blood vessel corresponding to that vascular group.

6. The multimodal fusion intelligent recognition and diagnostic assistance system for maternal and child medical images as described in claim 5, characterized in that, It also includes calculations based on the following formula: ； In the formula, express The average distance change rate of each vascular group in two adjacent diagnostic cycles; This indicates the total number of blood vessels in the group; Indicates the first The vascular group in the first The diagnostic cycle and the first - The change in distance over one diagnostic cycle; Indicates the first In the first diagnostic cycle The distance from the center of a vascular group to a point on the nearest blood vessel corresponding to that vascular group; Indicates the first -1 diagnostic cycle The distance from the center of a vascular group to a point on the nearest blood vessel corresponding to that vascular group.

7. The multimodal fusion intelligent recognition and diagnostic assistance system for maternal and child medical images as described in any one of claims 5 to 6, characterized in that, Based on the calculation results, if the distance between the blood vessel and the corresponding blood vessel group decreases, the system determines that the lesion at the gland is developing towards a worsening trend; otherwise, no determination is made.

8. The multimodal fusion intelligent recognition and diagnostic assistance system for maternal and child medical images as described in claim 7, characterized in that, If the system determines that the lesion at the gland is developing towards a worsening trend, it acquires the changes in the distance between the blood vessel and the corresponding blood vessel group based on three consecutive diagnostic cycles. If the blood vessels in the blood vessel group show a trend of dispersion, the system does not determine that the lesion at the gland is developing towards a worsening trend; otherwise, it does.

9. A method applied to the multimodal fusion intelligent recognition and diagnostic assistance system for maternal and child medical images as described in claim 1, characterized in that the steps... include: Acquire medical images of relevant parts of the patient, including areas corresponding to glandular tissue, and perform relevant processing on the medical images, including distinguishing the medical images into images corresponding to glands and images corresponding to connective tissue; The edges of glands and connective tissues are acquired from images corresponding to glands and images corresponding to connective tissues, and the clarity or blurriness of the edges of glands and connective tissues is determined. Feature extraction is performed on blurred edges in glands and connective tissue. The extraction steps include: Divide the edge of the gland and connective tissue into 6 to 10 edge points, and give a monitoring point based on each edge point. The monitoring point is the monitoring point that is closest to and clear to the corresponding edge point, and calculate the distance between the monitoring point and the corresponding edge point. Given the outline of each monitoring point, calculate the area of ​​each monitoring point; A preset diagnostic cycle is defined, wherein each diagnostic cycle consists of 7 to 10 days, and the change characteristics of the area are obtained in each diagnostic cycle. Based on the change characteristics, relevant information is obtained, including density, which is the degree of density of the monitoring points corresponding to the gland and the connective tissue when the area of ​​the monitoring points corresponding to each edge point of the gland and each edge point of the connective tissue is changing in a shrinking trend. The density levels are compared. The judgment is made based on the comparison results.

Images