CN-121661642-B - Image processing method for dynamic monitoring of cell proliferation

CN121661642BCN 121661642 BCN121661642 BCN 121661642BCN-121661642-B

Abstract

The invention relates to the technical field of image processing, in particular to an image processing method for dynamically monitoring cell proliferation. The method comprises the steps of obtaining an original microscopic image, preprocessing, constructing a three-channel tensor of a structure based on the preprocessed microscopic image, introducing a disturbance mechanism driven by a structural response to obtain a disturbance tensor of the preprocessed microscopic image, projecting the disturbance tensor of the preprocessed microscopic image to a combined space-frequency domain to obtain a multi-scale frequency response, superposing the multi-scale frequency response to obtain a space-frequency combined image response, carrying out brightness compression and edge smoothing on the space-frequency combined image response to obtain a reconstructed microscopic image, and carrying out cell tracking and behavior recognition to realize dynamic monitoring of cell proliferation. The problems that in the process of dynamic monitoring of cell proliferation, image processing is inaccurate, cell tracking is easy to break in a complex scene, and the behavior recognition capability lacks time sequence dynamic modeling capability are solved.

Inventors

ZHOU YANHENG
WANG JIALE
LIU YUE

Assignees

延安大学

Dates

Publication Date: 20260512
Application Date: 20260204

Claims (3)

1. An image processing method for dynamically monitoring cell proliferation is characterized by comprising the following steps: S1, acquiring an original microscopic image, preprocessing to obtain a preprocessed microscopic image, explicitly embedding brightness values of the preprocessed microscopic image and brightness gradients of the preprocessed microscopic image in the horizontal direction and the vertical direction into a three-dimensional tensor space to construct a three-channel tensor of a structure, introducing a disturbance mechanism driven by structural response, and constructing a disturbance response function by calculating second derivatives of the preprocessed microscopic image in the horizontal direction and the vertical direction and combining a nonlinear enhancement index: ; Wherein, the Is a disturbance response function representing the intensity of the structural response at the pixel location (x, y); is the brightness function of the preprocessed microscopic image; And Respectively representing second derivatives of the preprocessed microscopic image in the horizontal direction and the vertical direction; based on the disturbance response function, a disturbance energy modulation item is constructed, and a disturbance tensor of the preprocessed microscopic image is obtained by combining the three-channel tensor of the structure, wherein the disturbance tensor is specifically expressed as follows: ; Wherein, the The disturbance tensor of the preprocessed microscopic image is beta, and the beta is a sine disturbance amplitude control factor; is a disturbance energy modulation term; representing three channel tensors of the structure, wherein (x, y) is the two-dimensional space coordinate of the preprocessed microscopic image; Projecting the disturbance tensor of the preprocessed microscopic image to a combined space-frequency domain to obtain multi-scale frequency response; S2, introducing a disturbance energy weighting mechanism, applying double integration based on a disturbance response function, quantifying energy accumulation of the integral structure change intensity of the preprocessed microscopic image to obtain disturbance energy, calculating a fusion weight factor based on the disturbance energy, carrying out weighted superposition on multi-scale frequency response based on the fusion weight factor to obtain a space-frequency combined image response, carrying out brightness compression and edge smoothing on the space-frequency combined image response to obtain a reconstructed microscopic image, and carrying out cell tracking and behavior recognition based on the reconstructed microscopic image to realize dynamic monitoring of cell proliferation.
2. The method for processing images for dynamic monitoring of cell proliferation according to claim 1, wherein S1 specifically comprises: The method comprises the steps of constructing a two-dimensional frequency response kernel function by adopting a Gaussian-sinusoidal composite frequency kernel, and enabling the two-dimensional frequency response kernel function to act on a brightness channel of a disturbance tensor of a preprocessed microscopic image to obtain multi-scale frequency response, wherein the two-dimensional frequency response kernel function is constructed based on Gaussian attenuation items and sinusoidal modulation items, and the specific representation forms are as follows: , wherein, Is a gaussian decay term; is the frequency attenuation factor under the scale s, u and v are relative displacement and are used for describing the range of the two-dimensional frequency response kernel function definition domain; Is a sinusoidal modulation term; is a horizontal direction modulation factor; Is the vertical modulation factor at the scale s.
3. The method for processing images for dynamic monitoring of cell proliferation according to claim 1, wherein S2 specifically comprises: And (3) carrying out brightness compression and edge smoothing on the space-frequency combined image response by calculating the gradient amplitude of the space-frequency combined image response and introducing a gradient penalty coefficient and a nonlinear enhancement index of a power function to obtain a reconstructed microscopic image.

Description

Image processing method for dynamic monitoring of cell proliferation Technical Field The invention relates to the technical field of image processing, in particular to an image processing method for dynamically monitoring cell proliferation. Background With the development of life science and accurate medical technology, cell-level dynamic behavior observation and automated analysis have become important means for studying cell cycle, tumor evolution, drug response and intervention mechanisms. In the process, the microscopic imaging technology is used as a core supporting means, so that continuous imaging of cells in-vitro or in-vivo environment can be realized, and a rich data source is provided for subsequent image analysis. However, due to the limitations of the optical performance, illumination conditions, imaging time span and complexity of the tissue environment of the microscopic imaging system, the actually acquired cell image sequence often has the problems of uneven brightness, local blurring, lower signal-to-noise ratio, discontinuous structure, displacement between image frames and the like, and the accuracy and stability of subsequent image segmentation, cell tracking and behavior recognition are seriously affected. The existing image enhancement methods mostly adopt means such as histogram equalization, retinex theory, edge filtering or frequency domain enhancement, but the methods are often limited to single frame processing of images, so that the image quality is difficult to optimize from two dimensions of spatial structure and time continuity, and the problems of excessive enhancement or artifact generation exist in edge and texture detail enhancement. In the aspect of cell tracking, the existing method mostly depends on simple geometric matching or short-range corresponding strategies based on templates, and cannot process complex movements, cell shielding or division and other behaviors, so that the movement track of cells is interrupted or targets are confused. At the aspect of behavior recognition, the existing algorithm generally depends on a fixed rule or a shallow classifier, and lacks modeling capability for a complex evolution mode of cells in a time dimension. Disclosure of Invention The invention provides an image processing method for dynamic monitoring of cell proliferation, which aims to solve the technical problems that in the process of dynamic monitoring of cell proliferation, the image processing is inaccurate, the cell tracking is easy to be interrupted in a complex scene, and the behavior recognition capability lacks the time sequence dynamic modeling capability. The invention relates to an image processing method for dynamic monitoring of cell proliferation, which specifically comprises the following technical scheme: An image processing method for dynamically monitoring cell proliferation, comprising the following steps: S1, acquiring an original microscopic image, preprocessing to obtain a preprocessed microscopic image, constructing a three-channel tensor of a structure based on the preprocessed microscopic image, introducing a disturbance mechanism driven by a structural response, and constructing a disturbance response function to obtain a disturbance tensor of the preprocessed microscopic image; s2, introducing a disturbance energy weighting mechanism, superposing the multi-scale frequency response to obtain a space-frequency combined image response, carrying out brightness compression and edge smoothing on the space-frequency combined image response to obtain a reconstructed microscopic image, and carrying out cell tracking and behavior recognition based on the reconstructed microscopic image to realize dynamic monitoring of cell proliferation. Preferably, the S1 specifically includes: And explicitly embedding the brightness value of the preprocessed microscopic image and the brightness gradient of the preprocessed microscopic image in the horizontal direction and the vertical direction into a three-dimensional tensor space to construct a three-channel tensor of the structure. Preferably, the S1 specifically includes: In the implementation process of a disturbance mechanism driven by structural response, a disturbance response function is constructed by calculating second derivatives of the preprocessed microscopic image in the horizontal direction and the vertical direction and introducing a nonlinear enhancement index. Preferably, the S1 specifically includes: And constructing a disturbance energy modulation item based on a disturbance response function, and combining the three-channel tensor of the structure to obtain a disturbance tensor of the preprocessed microscopic image. Preferably, the S1 specifically includes: And constructing a two-dimensional frequency response kernel function by adopting a Gaussian-sinusoidal composite frequency kernel, and applying the two-dimensional frequency response kernel function to a brightness channel of a disturbance tens