Left: Slice of a 3D CT arthrography image showing the cartilage as dark lines between the bright cortical bone rims and contrast agent. Right: Corresponding anatomical section. Note that in the anatomical sections the cartilage layers appear bright in contrast to the dark appearance of cartilage in the CT images. The rectangular regions are the positions of thickness estimates between Radius and Lunate (upper rectangles) and between Lunate and Capitate (lower rectangles). The rectangular regions at the upper right corner of each image are magnifications of the ROI used for the thickness estimates.
Attenuation coefficient profile along a line perpendicular to the cartilage sheets (dashed lines) and the convolution of the attenuation coefficient profile with the PSF of the CT scanner (solid curve).
Estimation of cartilage layer orientation from image gradients in a ROI of pixels.
Cartilage-bone phantom for evaluation of the thickness estimation method. The layers in the phantom represent trabecular bone, cortical bone rims, cartilage, and contrast agent.
CT images of the cartilage-bone phantom at two orientations. Left: 0°, right: approx. 30°. The bright lines are cross sections of the Teflon sheets representing the cortical bone and the contrast agent. The dark lines represent the cartilage.
Simulation results showing the relationship between the real thickness of a cartilage layer and the thickness estimated from second order derivative zero crossings across the cartilage layers . FWHM of the PSF is (▴). FWHM of the PSF is (∎).
Fit of the model intensity profile to the CT image data of the cartilage-bone phantom.
Dependence of the thickness estimate on the in-plane orientation of the cartilage-bone phantom. The true diameter is as indicated by the horizontal dashed line.
Relationship between the real thickness of a cartilage layer and the estimated thickness as obtained from second order derivative zero crossings across the cartilage layers (∎) and our model-based method (•). The error bars indicate the standard deviation in the measurements. The standard deviations in the second derivative zero crossing method are extremely small and therefore not visible in the figure.
Comparison of cartilage thickness estimated from CT arthrography images (filled bars) and from anatomical sections (open bars). The cartilage thickness was estimated from layers situated between the Radius and the Lunate and between the Lunate and Capitate (square regions in Fig. 1.)
Article metrics loading...
Full text loading...