CN-121998880-A - Pulmonary nodule vision enhancement recognition method

CN121998880ACN 121998880 ACN121998880 ACN 121998880ACN-121998880-A

Abstract

The invention relates to the technical field of image data processing, in particular to a lung nodule visual enhancement identification method which comprises the steps of collecting lung CT gray level images, obtaining gray level ranges of the images as threshold selection ranges of image segmentation, performing image segmentation by taking traversal of different gray levels from small to large one by one as thresholds, calculating first membership probabilities of pixel points in each connected domain after each segmentation, counting the first membership probabilities of the same pixel point to obtain a probability change curve, obtaining relevant curve characteristics according to the probability change curve, obtaining second membership probabilities according to the relevant curve characteristics, and performing pseudo-color processing on the images by utilizing the second membership probabilities. The invention solves the problem of adhesion of the connected domain of single-threshold segmentation, and obtains better visual enhancement effect of the pulmonary nodule region.

Inventors

ZHANG YU
YANG LIMIN
LIU LU
WEI WEI

Assignees

无锡市第九人民医院

Dates

Publication Date: 20260508
Application Date: 20230509

Claims (8)

1. A method for visually enhancing and identifying pulmonary nodules, the method comprising the steps of: Acquiring a CT gray level image of the lung by using a CT scanner, and acquiring a threshold selection range in the CT image according to the gray level range of the CT gray level image of the lung; Traversing the threshold value one by one according to the threshold value selection range, segmenting the lung CT gray level image by taking the threshold value as an image segmentation threshold value one by one to obtain a plurality of image segmentation result images, and obtaining the circumference and the area of a connected domain where the same pixel point is located in the result images; Obtaining first membership probability of a pixel point in each result image according to the perimeter and the area of a connected domain where the pixel point is located, forming a first membership probability sequence by the first membership probability of the same pixel point in a plurality of result images, taking a threshold value used when dividing an image threshold value as a horizontal axis, taking the first membership probability as a vertical axis, and constructing the first membership probability sequence into a change curve of the first membership probability relative to the threshold value, namely a first curve; Intercepting a second curve on the first curve, obtaining a maximum value point and a minimum value point of the first membership probability according to the second curve, obtaining a first curve characteristic of the second curve according to the difference between the maximum value point and the minimum value point, and obtaining a second curve characteristic of the second curve according to the slopes of the maximum value point and the minimum value point; Obtaining a second membership probability of the pixel point according to the first curve characteristic and the second curve characteristic; and performing gamma conversion on the lung CT gray-scale image according to the second membership probability of the pixel point to obtain a gamma conversion result image, and performing pseudo-color processing on the gamma conversion result image to obtain a pseudo-color processing result image.
2. The method for identifying the pulmonary nodule vision enhancement according to claim 1, wherein the step of obtaining the threshold selection range in the CT image according to the gray scale range of the CT image of the lung comprises the following specific steps: The method comprises the steps of obtaining a CT gray level image of a lung by using a CT scanner, counting the number of pixels with the same gray level value in the image, establishing a gray level distribution histogram by taking the gray level value as a horizontal axis and the corresponding number of pixels as a vertical axis, obtaining a maximum gray level value and a minimum gray level value of the corresponding CT gray level image of the lung according to the gray level distribution histogram, obtaining a gray level range of the CT gray level image of the lung within a range of the maximum gray level value and the minimum gray level value, and taking the gray level value within the gray level range of the CT gray level image of the lung as a threshold value selection range.
3. The method of claim 1, wherein the resulting image is pre-processed using morphological open operations.
4. The method for identifying pulmonary nodule visual enhancement according to claim 1, wherein the first sequence of membership probabilities is obtained by: first membership probability of pixel points in lung CT gray scale image: Wherein A in refers to the area of the connected domain region where the ith pixel point is located in the nth result image, C in refers to the perimeter of the connected domain where the ith pixel point is located in the nth morphological processing result image, and m in refers to the first membership probability of the ith pixel point in the nth result image; the first membership probability of the pixel point in all the result images forms a first membership probability sequence.
5. The method for identifying pulmonary nodule vision enhancement of claim 1, wherein the first curve feature is obtained by: Wherein f i represents the fluctuation degree of the second curve of the ith pixel point in the lung CT gray scale image, which is marked as a first curve characteristic, H represents the number of first membership probabilities greater than the empirical probability threshold in the first membership probability sequence of the ith pixel point in the lung CT gray scale image, M ih represents the H first membership probabilities greater than the empirical probability threshold in the first membership probability sequence of the ith pixel point in the lung CT gray scale image, and the local mean factor Representing the mean of the first membership probabilities in the second curve.
6. The method for identifying pulmonary nodule vision enhancement of claim 1, wherein the second curve features are obtained by: Obtaining a pair of maximum value points and a pair of minimum value points which correspond to the maximum difference on the abscissa in the second curve; The difference value between the maximum value and the minimum value of the first membership probability in the second curve is recorded as a first difference value; Recording the difference value of the segmentation threshold corresponding to the maximum value of the first membership probability and the target minimum value in the second curve as a second difference value; the absolute value of the ratio of the first difference and the second difference is taken as the second curve characteristic.
7. The method for identifying pulmonary nodule vision enhancement according to claim 1, wherein the step of obtaining the second membership probability of the pixel point according to the first curve feature and the second curve feature comprises the following specific steps: P i ＝exp(-βf i )*exp(-k i ) Wherein exp () is an exponential function based on a natural constant, beta is an adjustment parameter, P i represents a second membership probability of an ith pixel point in the lung CT gray scale image, a first curve characteristic f i represents a fluctuation degree of a second curve of the ith pixel point in the lung CT gray scale image, and a second curve characteristic k i represents a slope of the second curve of the ith pixel point in the lung CT gray scale image.
8. The method for identifying pulmonary nodule vision enhancement according to claim 1, wherein the step of performing gamma transformation on the pulmonary CT gray scale image according to the second membership probability of the pixel to obtain a gamma transformation result image comprises the following specific steps: Wherein q i represents the gray value of the ith pixel in the image, P i represents the second membership probability of the ith pixel, f i represents the gray value of the ith pixel after gamma conversion, and all the pixels after gamma conversion form a gamma conversion result image.

Description

Pulmonary nodule vision enhancement recognition method Technical Field The invention relates to the technical field of image data processing, in particular to a pulmonary nodule visual enhancement recognition method. Background Lung CT is an imaging examination method, a technique for determining chest lung disease by computed tomography. This technique can be used to diagnose pulmonary diseases such as pulmonary nodules, lung tumors, lung inflammation in lung tissue lesions. The medical department judges whether the lung has a nodule through the lung CT image, firstly, the lung CT gray level image is required to be preprocessed, then the lung region is extracted, and then the lung region is segmented, and the probability of the nodule is judged through the shape and the area of the nodule. In the prior art, an adaptive threshold segmentation method is adopted, but in the result of the adaptive threshold segmentation, due to the adhesion problem of a connected domain, a node and corresponding tissues are connected and integrated, and the subsequent judgment of the shape and the area of the node is greatly interfered. The method obtains the lung region by firstly carrying out segmentation treatment on the image. The threshold is approximately set by the lung region gray level histogram. And analyzing the probability of the nodule under each threshold according to the characteristics of the nodule, and acquiring the probability of the nodule in the CT image. And finally, carrying out pseudo-color enhancement on the obtained nodule. Disclosure of Invention The invention provides a pulmonary nodule visual enhancement recognition method, which aims to solve the existing problems. The invention relates to a pulmonary nodule vision enhancement recognition method which adopts the following technical scheme: one embodiment of the present invention provides a method for visually enhanced identification of pulmonary nodules, the method comprising the steps of: Acquiring a CT gray level image of the lung by using a CT scanner, and acquiring a threshold selection range in the CT image according to the gray level range of the CT gray level image of the lung; Traversing the threshold value one by one according to the threshold value selection range, segmenting the lung CT gray level image by taking the threshold value as an image segmentation threshold value one by one to obtain a plurality of image segmentation result images, and obtaining the circumference and the area of a connected domain where the same pixel point is located in the result images; Obtaining first membership probability of a pixel point in each result image according to the perimeter and the area of a connected domain where the pixel point is located, forming a first membership probability sequence by the first membership probability of the same pixel point in a plurality of result images, taking a threshold value used when dividing an image threshold value as a horizontal axis, taking the first membership probability as a vertical axis, and constructing the first membership probability sequence into a change curve of the first membership probability relative to the threshold value, namely a first curve; Intercepting a second curve on the first curve, obtaining a maximum value point and a minimum value point of the first membership probability according to the second curve, obtaining a first curve characteristic of the second curve according to the difference between the maximum value point and the minimum value point, and obtaining a second curve characteristic of the second curve according to the slopes of the maximum value point and the minimum value point; Obtaining a second membership probability of the pixel point according to the first curve characteristic and the second curve characteristic; and performing gamma conversion on the lung CT gray-scale image according to the second membership probability of the pixel point to obtain a gamma conversion result image, and performing pseudo-color processing on the gamma conversion result image to obtain a pseudo-color processing result image. Further, the step of obtaining the threshold selection range in the CT image according to the gray scale range of the lung CT gray scale image comprises the following specific steps: The method comprises the steps of obtaining a CT gray level image of a lung by using a CT scanner, counting the number of pixels with the same gray level value in the image, establishing a gray level distribution histogram by taking the gray level value as a horizontal axis and the corresponding number of pixels as a vertical axis, obtaining a maximum gray level value and a minimum gray level value of the corresponding CT gray level image of the lung according to the gray level distribution histogram, obtaining a gray level range of the CT gray level image of the lung within a range of the maximum gray level value and the minimum gray level value, and taking the gray level value withi