CN-122024234-A - Method and system for detecting quality of medical image of embryo in cleavage stage

CN122024234ACN 122024234 ACN122024234 ACN 122024234ACN-122024234-A

Abstract

The invention discloses a method and a system for detecting the medical image quality of a split-phase embryo, which relate to the technical field of medical image detection and comprise the following steps that S1, the split-phase embryo is placed in an adaptive agarose culture dish, and the refractive index of the agarose culture dish is consistent with that of embryo culture solution; S2, constructing a 4D imaging model, continuously and dynamically imaging the embryo in the cleavage period, collecting 4D original image data of the embryo in the cleavage period within a preset time range, S3, constructing a pretreatment model, carrying out refraction correction on the 4D original image data, identifying embryo contours and cell distribution, judging embryo body positions, forming pretreated image data, and carrying out light refraction deviation compensation by adopting an agarose culture dish with the same refractive index as that of an embryo culture solution and matching with a special refraction correction submodule, so that the refraction deviation is eliminated from the source and the later treatment double dimensions, the problems of image blurring and distortion are effectively solved, and the basic quality of the image is improved.

Inventors

TUO YA
WANG WEI
XUAN XIAOYUE
YANG CAIYUN

Assignees

沈阳升维医疗科技有限公司
内蒙古医科大学附属医院(内蒙古自治区心血管研究所)

Dates

Publication Date: 20260512
Application Date: 20260411

Claims (10)

1. The method for detecting the quality of the medical image of the embryo in the cleavage stage is characterized by comprising the following steps of: S1, placing an embryo in an cleavage stage in an adaptive agarose culture dish, wherein the refractive index of the agarose culture dish is consistent with that of an embryo culture solution; S2, constructing a 4D imaging model, continuously and dynamically imaging the embryo in the cleavage period, and collecting 4D original image data of the embryo in the cleavage period within a preset time range; s3, constructing a preprocessing model, carrying out refraction correction on the 4D original image data, identifying embryo contours and cell distribution, judging embryo body positions, and forming preprocessed image data; s4, performing multi-scale enhancement processing and artifact removal optimization processing on the preprocessed image data; S5, performing subcellular structure dynamic capturing and analysis on the optimized image data; And S6, carrying out quality screening on the analyzed image data, outputting a visual embryo image quality detection report, and simultaneously marking a specific defect type on the unqualified embryo image.
2. The method for detecting the quality of medical images of embryos in the cleavage stage according to claim 1, wherein in S2, the specific mathematical formula of the 4D imaging model is: Wherein: For spatial coordinates in 4D imaging An image pixel intensity value at time t; Initial light intensity for incident light; in space coordinates for embryonic tissue Light absorption coefficient at; A propagation path length for incident light within the embryonic tissue; the dynamic changes of the morphology and distribution of embryo cells along with time are reflected for the light modulation coefficient of embryo tissues; Is a random noise term of the imaging system.
3. The method for detecting the medical image quality of the embryo at the cleavage stage according to claim 1, wherein in the step S2, the preset time range is 12-48 h of a key period of embryo development at the cleavage stage, the imaging interval is 2-30 min, and the imaging resolution is 300nm-0.5 μm.
4. The method for detecting the quality of the medical image of the embryo at the cleavage stage according to claim 1, wherein in the step S3, the preprocessing model integrates refraction correction, contour recognition, cell distribution analysis and body position judgment sub-modules, the refraction correction sub-modules are used for carrying out light refraction deviation compensation on the 4D original image data, the contour recognition and cell distribution analysis sub-modules are used for completing feature extraction of embryo areas and cells, and finally the body position judgment sub-modules are used for judging the embryo body position based on the cell space distribution features.
5. The method for detecting the quality of medical images of embryos in the cleavage period according to claim 4, wherein the specific mathematical formula of the refraction correction sub-module is as follows: Wherein: Is the space coordinate after refraction correction Image pixel intensity values at time t; the pixel light intensity value of the 4D original image; The standard refractive index of the embryo culture solution and the agarose culture dish; in space coordinates for embryonic tissue An actual refractive index at; The refractive index correction coefficient is a value range of 0.95-1.05; Is the optical path difference between the incident light and the culture solution passing through the embryo tissue.
6. The method for detecting the quality of the medical image of the embryo at the cleavage stage according to claim 4, wherein the specific formula of the contour recognition and cell distribution analysis submodule is as follows: Wherein: Is space coordinates Pixel edge gradient values at time t; respectively the partial conductance gradient values of three space dimensions, Reflecting the space coordinates for the light intensity value of the image pixel after refraction correction The small variation of the light intensity of the image pixels after correction at time t, Characterizing the tiny distance variation in the corresponding space dimension as partial differentiation of X, Y, Z axis coordinates in the three-dimensional space Greater than a preset gradient threshold When the pixel is judged to be an embryo contour pixel; The body position judging submodule is characterized by the spatial arrangement gravity centers of embryo blastomeres and cell stacking angles, judges that the embryos are normal, lateral and inverted, judges that the embryos are normal when the gravity center offset is smaller than 1/5 of the maximum outline diameter of the embryos, and judges that the embryos are lateral or inverted according to the stacking angles otherwise.
7. The method for detecting the quality of the medical image of the embryo in the cleavage stage according to claim 1, wherein in S4, the multiscale enhancement processing adopts a gaussian pyramid multiscale decomposition algorithm, and the specific formula is as follows: Wherein: As a function of the gaussian kernel, The scale factor is a value range of 0.5-2.0, ∗ is convolution operation; Is an image after single-scale Gaussian filtration; The value of the decomposition scale number is 3-5; For each scale weight coefficient, satisfy ; Is the image data after multi-scale enhancement.
8. The method for detecting quality of medical images of embryo at cleavage stage according to claim 1, wherein in S4, the artifact removal optimization process adopts a three-dimensional wiener filtering algorithm to remove motion artifacts and system artifacts in an imaging process, and the specific formula is as follows: Wherein: Image data optimized for artifact removal; Is a transfer function of the imaging system; the noise coefficient is 0.01-0.1; Is a modulo operation.
9. The method for detecting the quality of the medical image of the embryo in the cleavage stage according to claim 1, wherein in S5, the dynamic capturing and analysis of the subcellular structure are realized by constructing a subcellular feature detection model, the feature points of the subcellular structure are extracted from the optimized image data, and then the dynamic track tracking and the parameter analysis of the time sequence are performed on the feature points, and the specific formula of the subcellular feature point extraction is as follows: Wherein: Is the response value of subcellular structure characteristic points; Is a Laplacian operator; Is a gradient operator; Is a norm operation; Is a minimum constant, takes a value of 10 < -6 > and avoids the denominator of 0 when And when the characteristic point is larger than the preset characteristic threshold value, judging the characteristic point as a subcellular structure characteristic point.
10. A system for detecting the quality of medical images of embryos in the cleavage stage, which is used for realizing the method for detecting the quality of medical images of embryos in the cleavage stage, which is characterized by comprising any one of claims 1 to 9.

Description

Method and system for detecting quality of medical image of embryo in cleavage stage Technical Field The invention relates to the technical field of medical image detection, in particular to a method and a system for detecting the quality of medical images of embryos in the cleavage stage. Background Medical image detection of embryo in cleavage stage is a key technical link in auxiliary reproduction field, and the quality of embryo image directly determines the accuracy of embryo development state evaluation and high-quality embryo screening. In the current embryo imaging detection process in the cleavage stage, a plurality of technical pain points exist, namely, on one hand, the difference of refractive indexes of embryo culture solution and a culture dish is easy to cause imaging light refraction deviation to cause image blurring and contour distortion, on the other hand, the traditional imaging is mostly static or low-dimensional dynamic imaging, cell dynamic changes in a key development period of 12h-48h of embryos in the cleavage stage cannot be completely captured, imaging resolution is insufficient, subcellular structures are difficult to clearly present, in addition, motion artifacts and system artifacts are easy to generate in the imaging process, image preprocessing is only simple and filtering, targeted multi-scale enhancement and accurate artifact removal processing are lacked, meanwhile, dynamic analysis capability of the subcellular structures is lacked, and finally, the embryo image quality evaluation lacks quantitative basis, and the screening result is strong in subjectivity and low in accuracy. Disclosure of Invention The embodiment of the invention provides a medical image quality detection method and a medical image quality detection system for a split-phase embryo, which aim to solve the problems that light refraction deviation cannot be completely captured, cell dynamic changes in a critical development period of 12-48 h of the split-phase embryo are lacked, targeted multi-scale enhancement is lacked, accurate artifact removal treatment is carried out, and dynamic analysis capability of subcellular structures is lacked. In order to achieve the above purpose, the embodiment of the present invention adopts the following technical scheme: a method for detecting the quality of medical images of embryos in the cleavage stage comprises the following steps: S1, placing an embryo in an cleavage stage in an adaptive agarose culture dish, wherein the refractive index of the agarose culture dish is consistent with that of an embryo culture solution; S2, constructing a 4D imaging model, continuously and dynamically imaging the embryo in the cleavage period, and collecting 4D original image data of the embryo in the cleavage period within a preset time range; s3, constructing a preprocessing model, carrying out refraction correction on the 4D original image data, identifying embryo contours and cell distribution, judging embryo body positions, and forming preprocessed image data; s4, performing multi-scale enhancement processing and artifact removal optimization processing on the preprocessed image data; S5, performing subcellular structure dynamic capturing and analysis on the optimized image data; And S6, carrying out quality screening on the analyzed image data, outputting a visual embryo image quality detection report, and simultaneously marking a specific defect type on the unqualified embryo image. Further, in S2, the specific mathematical formula of the 4D imaging model is: Wherein: For spatial coordinates in 4D imaging An image pixel intensity value at time t; Initial light intensity for incident light; in space coordinates for embryonic tissue Light absorption coefficient at; A propagation path length for incident light within the embryonic tissue; the dynamic changes of the morphology and distribution of embryo cells along with time are reflected for the light modulation coefficient of embryo tissues; Is a random noise term of the imaging system. Further, in S2, the preset time range is 12-48 hours of the critical period of embryo development in the cleavage stage, the imaging interval is 2-30 min, and the imaging resolution reaches 300nm-0.5 μm. Further, in S3, the preprocessing model integrates refraction correction, contour recognition, cell distribution analysis and body position judgment sub-modules, the refraction correction sub-modules are used for performing light refraction deviation compensation on the 4D original image data, then the contour recognition and cell distribution analysis sub-modules are used for completing feature extraction of embryo areas and cells, and finally the body position judgment sub-modules are used for judging embryo body positions based on cell space distribution features. Further, the specific mathematical formula of the refraction correction sub-module is as follows: Wherein: Is the space coordinate after refraction correction Image pixel inten