The intensity in CT images is associated to tissue density and it is measured with the Hounsfield scale (HS). Specific HS value ranges are defined for each type of tissue, which allows the direct comparison of images from different vendors. However, artifacts or acquisition errors might affect the image and give pixel values outside the physiological range. We filter these artifacts by thresholding the pixel values of CT images between HS=−1050 (air density score) and HS=3050 (bone density score). Similarly, the SUVs in PET images are thresholded in the range between 0 and 50 to avoid artifacts due to errors in sensors readings. 解释

这段话主要讲了CT和PET图像中像素值的测量和过滤。在CT图像中，像素值与组织密度相关，使用的是Hounsfield刻度（HS），每种类型的组织都有特定的HS值范围。但是，可能会出现伪影或采集误差，导致像素值超出生理范围。为了过滤这些伪影，需要将CT图像的像素值限定在HS=-1050（空气密度）和HS=3050（骨密度）之间。同样地，在PET图像中，SUV值也需要进行限制，通常在0到50之间，以避免由于传感器读数误差导致的伪影。

相关问题

Implement a function that equalizes grey-scale images based on their histogram. The input is a given grey level image; the output is the derived image with uniform intensity distribution.

Sure, I can help you with that! Here is a Python implementation of the function:

import numpy as np
import cv2

def hist_equalization(img):
    # Calculate histogram
    hist, bins = np.histogram(img.flatten(), 256, [0, 256])
    # Calculate cumulative distribution function (CDF)
    cdf = hist.cumsum()
    # Normalize CDF
    cdf_normalized = cdf * hist.max() / cdf.max()
    # Create lookup table
    lut = np.interp(np.arange(256), bins[:-1], cdf_normalized)
    # Apply lookup table to image
    equalized_img = cv2.LUT(img, lut)
    return equalized_img

To use this function, simply pass in a grayscale image as the argument and it will return the equalized image. Here's an example of how to use it:

# Load grayscale image
img = cv2.imread('image.jpg', 0)
# Apply histogram equalization
equalized_img = hist_equalization(img)
# Display original and equalized images
cv2.imshow('Original Image', img)
cv2.imshow('Equalized Image', equalized_img)
cv2.waitKey(0)
cv2.destroyAllWindows()

I hope this helps! Let me know if you have any questions.

Implement a function with Matlab that equalizes grey-scale images based on their histogram. The input is a given grey level image; the output is the derived image with uniform intensity distribution.

Sure, here's an implementation of the function in MATLAB:

function equalized_img = hist_equalization(img)
    % Calculate histogram
    [hist, bins] = imhist(img);
    % Calculate cumulative distribution function (CDF)
    cdf = cumsum(hist);
    % Normalize CDF
    cdf_normalized = cdf * max(hist) / max(cdf);
    % Create lookup table
    lut = interp1(bins, cdf_normalized, 0:255);
    % Apply lookup table to image
    equalized_img = uint8(lut(double(img)+1));
end

To use this function, simply pass in a grayscale image as the argument and it will return the equalized image. Here's an example of how to use it:

% Load grayscale image
img = imread('image.jpg');
img_gray = rgb2gray(img);
% Apply histogram equalization
equalized_img = hist_equalization(img_gray);
% Display original and equalized images
subplot(1,2,1), imshow(img_gray), title('Original Image')
subplot(1,2,2), imshow(equalized_img), title('Equalized Image')

I hope this helps! Let me know if you have any questions.