4f) and the leg portion of the mask ( Fig. Both the resulting two class image ( Fig. The estimated classification threshold was then reduced by 50% to account for intensity inhomogeneity in the image. 3a) using a two-class global histogram based intensity thresholding method ( Fig. The muscle region was identified on the water-only image derived from the mDixon scan ( Fig. Then, the produced image was overlaid on the sodium image ( Fig.
#Segmentation imuscle skin
The reduced skin region was parallel to the coil surface and the tissue thickness was more stable. Therefore, the skin region was reduced to include only the portion in contact with the surface of the phantom holder ( Fig. At the time of MR acquisition, the posterior area of the leg was resting on the phantom holder surface and thus aligned perpendicular to the slice direction such that through-plane partial volume effect was minimized. It should be noted that this process assumes the skin thickness is similar in all participants. 4b) to select approximately the outer 4 mm of the leg region ( Fig. 4b) by a 4 mm radius circular kernel and subtracting the resulting image from the original leg portion of the mask ( Fig. The skin region was estimated by eroding the leg portion of the mask ( Fig. Using the sodium coil and an optimized 3D gradient-echo sequence, a sodium image was obtained with the following parameters: FOV = 192 x 192 x 210 mm 3, voxel size = 3 x 3 x 30 mm 3, 7 slices, TR/TE/FA = 130 ms/0.99 ms/90°, bandwidth = 434 Hz/pixel, acquisition: 26, and scan time = 15 min 54 s. The standard proton-density-weighted scan used the following parameters: TR/TE/FA = 4000 ms/ 30 ms/ 90°, and scan time = 2 min 32 s. The proton mDixon scan was acquired with TR = 200 ms and TE = 4.6 ms, 20 images were constructed in the form of water, water fat in-phase, water fat out-of-phase, and fat images, scan time = 3 min 52 s. These proton scans have the same geometry parameters: FOV = 192 x 192 mm 2, resolution = 1 x 1 mm 2, 5 slices at a thickness of 6 mm. Two proton scans were performed using the scanner body coil: a mDixon scan for fat and water images, and a standard proton-density-weighted image. The left lower leg was placed in the coil, in close proximity to a set of calibration phantoms (NaCl aqueous solution of 10mM, 20mM, 30mM, and 40mM). MR images were acquired on a Philips Achieva 3.0T MR scanner (Philips Healthcare, Cleveland OH, USA) using a 23Na quadrature knee coil (Rapid Biomedical GmbH, Rimpar, Germany).
#Segmentation imuscle manual
In this study, we develop an application-specific automated segmentation pipeline for the lower leg and show that its segmentations applied to sodium MR images yield sodium concentration measurements comparable to the measurements obtained via the gold standard manual segmentations. For segmentation, other studies use a combination of applications including: shaping histograms, adaptive thresholding, connectivity, a deformable model, global histogram based intensity thresholding, k means clustering, and intensity based temporal homomorphic filter. One approach has been to apply a fuzzy clustering method to segment anatomical regions such as adipose tissue, cortical bone, and spongy bone in the lower musculature of the leg and in the thigh. A variety of automated methodologies have been developed to segment anatomical magnetic resonance images of the leg. This process requires meticulous attention and inconsistencies can be introduced via human error, which diminishes reproducibility.Īn automated approach could address these problems, and consequently improve workflow. Ī straightforward approach to determine sodium levels based on the sodium magnetic resonance images of the calf is to manually segment the desired regions. The lower leg muscle and skin is of particular interest because of the technical simplicity and speed of obtaining an MRI scan of the calf. With technological advancements, biomedical application of sodium (Na) magnetic resonance imaging is on the rise, as it provides unique and quantitative biochemical information related to tissue viability, cell integrity and function.
0 kommentar(er)
