Search

CN-122023822-A - Boundary recognition method, device, equipment and storage medium for physiological tissue

CN122023822ACN 122023822 ACN122023822 ACN 122023822ACN-122023822-A

Abstract

The application discloses a method, a device, equipment and a storage medium for identifying the boundary of a physiological tissue, wherein the method comprises the steps of extracting a boundary buffer area from a first image sequence of a target physiological tissue; registering the boundary buffer areas of at least two first image sequences to obtain a concerned area in the boundary buffer areas, and processing the concerned area based on the second image sequences of the target physiological tissues to obtain the target boundary of the target physiological tissues. According to the application, the region needing to be focused is determined through the boundary buffer regions of the two first image sequences, and the region needing to be focused is optimized depending on the second image sequence, so that the recognition accuracy of the tissue boundary can be effectively improved.

Inventors

  • XIAO YUETING
  • YANG GUANG
  • ZHENG CHAO
  • MAO XINSHENG

Assignees

  • 数坤科技股份有限公司

Dates

Publication Date
20260512
Application Date
20251231

Claims (10)

  1. 1. A method for identifying a boundary of a physiological tissue, comprising: Extracting a boundary buffer area from a first image sequence of a target physiological tissue, wherein the boundary buffer area comprises an initial boundary and a buffer area of the initial boundary; registering boundary buffer areas of at least two first image sequences to obtain a concerned area in the boundary buffer areas; And processing the concerned region based on the second image sequence of the target physiological tissue to obtain a target boundary of the target physiological tissue.
  2. 2. The method of claim 1, wherein the processing the region of interest based on the second image sequence of the target tissue to obtain the target boundary of the target tissue comprises: Preprocessing the first image sequence to obtain a first optimized image sequence which can be registered with the second image sequence; and inputting the first optimized image sequence marked with the region of interest and the second image sequence into a trained neural network model so as to process the region of interest through the neural network model and obtain a target boundary of the target physiological tissue.
  3. 3. The method of claim 2, wherein the preprocessing the first image sequence to obtain a first optimized image sequence that can be registered with the second image sequence comprises: Extracting tissue segmentation data in the first image sequence; Determining a tissue centerline of the target tissue based on the tissue segmentation data, and determining a plurality of centerline points on the tissue centerline; And straightening the target physiological tissue in the first image sequence based on the central line point of the tissue central line to obtain a first optimized image sequence which can be registered with the second image sequence.
  4. 4. The method of claim 1, wherein registering the boundary buffer areas of the at least two first image sequences to obtain the region of interest in the boundary buffer areas comprises: Performing feature matching on boundary buffer areas of at least two first image sequences, and determining target pixel points with feature differences in the boundary buffer areas; weighting the difference value of the target pixel point based on the weight associated with the target pixel point to obtain deviation information of each region in the boundary buffer region, wherein the weight associated with the target pixel point is determined based on the distance between the target pixel point and the initial boundary; And determining a region in which the deviation information exceeds a preset deviation threshold as a region of interest in the boundary buffer region.
  5. 5. The method of claims 1-4, wherein the first image sequence is a magnetic resonance imaging sequence and the second image sequence is a computed tomography imaging sequence.
  6. 6. The method of claim 5, wherein the first image sequence further comprises a black blood sequence, and wherein the processing the region of interest based on the second image sequence of the target tissue to obtain the target boundary of the target tissue comprises: Determining a constraint boundary of the target physiological tissue according to a boundary buffer area in a multi-frame black blood sequence in the first image sequence; and based on the second image sequence, using the constraint boundary as a boundary constraint condition of the target physiological tissue, and processing the concerned region in the boundary buffer region to obtain a target boundary of the target physiological tissue.
  7. 7. The method of claim 1, wherein the tissue comprises vascular tissue, and wherein the method further comprises: and determining fractional flow reserve of the target physiological tissue according to the target boundary of the target physiological tissue and the focus boundary of the focus in the target physiological tissue.
  8. 8. A boundary recognition device for a physiological tissue, comprising: The identifying module is used for extracting a boundary buffer area from a first image sequence of a target physiological tissue, wherein the boundary buffer area comprises an initial boundary and a buffer area of the initial boundary; the registration module is used for registering boundary buffer areas of at least two first image sequences to obtain a concerned area in the boundary buffer areas; And the processing module is used for processing the region of interest based on the second image sequence of the target physiological tissue to obtain a target boundary of the target physiological tissue.
  9. 9. A computer device, the computer device comprising: one or more processors; Memory, and One or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the processor to implement the boundary identification method of the physiological tissue of any one of claims 1 to 7.
  10. 10. A computer-readable storage medium, characterized in that a computer program is stored thereon, which computer program is loaded by a processor to perform the boundary recognition method of a physiological tissue according to any one of claims 1 to 7.

Description

Boundary recognition method, device, equipment and storage medium for physiological tissue Technical Field The application relates to the technical field of medical image processing, in particular to a method, a device, equipment and a storage medium for identifying boundaries of physiological tissues. Background Fractional flow reserve (Fractional Flow Reserve, FFR) is defined as the ratio of the maximum blood flow that a diseased vessel can provide to the maximum blood flow that the vessel can provide when it is completely normal. The index can be used for judging the ischemia degree of the lesion blood vessel, thereby accurately predicting the factors causing cardiovascular and cerebrovascular diseases such as heart and brain blood supply deficiency and the like. However, accurate calculation of fractional flow reserve is required to rely on clearer medical images of physiological tissue, as well as accurate medical images of lesion location. However, due to the real-time beating of the heart, the acquired medical images of the physiological tissue may have the problem of unclear boundaries, which results in that the clear outline boundaries and positions of the lesions cannot be determined, thereby affecting the accuracy of the subsequent calculation of the fractional flow reserve. Disclosure of Invention The application provides a boundary identification method, device, equipment and storage medium for physiological tissues, which at least partially improve the boundary identification precision of the physiological tissues so as to improve the calculation effect of fractional flow reserve. In a first aspect, the present application provides a method for identifying a boundary of a physiological tissue, comprising: Extracting a boundary buffer area from a first image sequence of a target physiological tissue, wherein the boundary buffer area comprises an initial boundary and a buffer area of the initial boundary; registering boundary buffer areas of at least two first image sequences to obtain a concerned area in the boundary buffer areas; And processing the concerned region based on the second image sequence of the target physiological tissue to obtain a target boundary of the target physiological tissue. In one embodiment of the present application, the processing the region of interest based on the second image sequence of the target physiological tissue to obtain a target boundary of the target physiological tissue includes: Preprocessing the first image sequence to obtain a first optimized image sequence which can be registered with the second image sequence; and inputting the first optimized image sequence marked with the region of interest and the second image sequence into a trained neural network model so as to process the region of interest through the neural network model and obtain a target boundary of the target physiological tissue. In one embodiment of the present application, the preprocessing the first image sequence to obtain a first optimized image sequence that can be registered with the second image sequence includes: Extracting tissue segmentation data in the first image sequence; Determining a tissue centerline of the target tissue based on the tissue segmentation data, and determining a plurality of centerline points on the tissue centerline; And straightening the target physiological tissue in the first image sequence based on the central line point of the tissue central line to obtain a first optimized image sequence which can be registered with the second image sequence. In one embodiment of the present application, the registering the boundary buffer areas of at least two first image sequences to obtain the region of interest in the boundary buffer areas includes: Performing feature matching on boundary buffer areas of at least two first image sequences, and determining target pixel points with feature differences in the boundary buffer areas; weighting the difference value of the target pixel point based on the weight associated with the target pixel point to obtain deviation information of each region in the boundary buffer region, wherein the weight associated with the target pixel point is determined based on the distance between the target pixel point and the initial boundary; And determining a region in which the deviation information exceeds a preset deviation threshold as a region of interest in the boundary buffer region. In one embodiment of the application, the first image sequence is a magnetic resonance imaging sequence and the second image sequence is a computed tomography imaging sequence. In one embodiment of the present application, the first image sequence further includes a black blood sequence, and the processing the region of interest based on the second image sequence of the target physiological tissue to obtain a target boundary of the target physiological tissue includes: Determining a constraint boundary of the target physiological tissue according to