In medical imaging or radiation physics, a phantom is an artificial object (a model) that mimics the physical or radiological properties of real biological tissue — like human organs, bone, or soft tissue.
A typical ion-beam radiograph gives only a 2D projection and you don’t know the depth of changes (i.e., how far into the beam the changed tissue is).
The team developed a method to scan through different “depths” virtually in the image and choose the one where the change looks sharpest. They propose a “2.5D” method to retrieve some depth information.
Water-equivalent thickness WET was calculated. It tells how much a material (like tissue) slows down an ion beam, measured as if it were a certain thickness of water. Denser or thicker tissue → higher WET. If the patient (or the phantom) changes like an air cavity appears, or tissue swells — the WET values change at those pixels.
Tenengrad focus‐measure method (image sharpness metric): This method measures a focus/sharpness metric based on image gradients which was computed using Sobel operators.
M is gradient magnitude at each pixel.
S is a sobel operator: simple image processing filter used to detect edges — places in an image where brightness (intensity) changes quickly. It calculates gradients — how steeply the pixel brightness changes in the x and y directions.
Tenengrad score is computed by calculating the average gradient magnitude over the entire image.Tenengrad measures overall sharpness of the image. it shows which depth inside the phantom the reconstructed image is sharpest.
Source: https://iopscience.iop.org/article/10.1088/1361-6560/ae02de#pmbae02des5
0 comments:
Post a Comment