Noise Filtering Methods in MRI - Advance Image Processing in Magnetic Resonance Imaging - page 176

15

10

5

0

∇ I 1

− 5

∇ I 2

− 10

∇ I 3

15

−

0 5 10 15 20 25 30

K

FIGURE 6.5 Plot of the slope increment (∆ S ) as a function of K for ∆ I 1 = 10, ∆ I 2 = 20,
and ∆ I 3 = 30.

The performance of anisotropic filtering in reducing noise while preserving
image structure was compared to results obtained at different NEX values.
Two sets of images were derived: one set obtained from a two-cylinder phantom
(a 60-mm-diameter cylinder filled with olive oil and immersed in a 100-mm-diameter
water-filled cylinder), and one set of images derived from a normal human heart.
Cardiac images were obtained from a group of five normal volunteers.
The series of MR images was acquired with a 1.5T GE Signa CV/i scanner
with a fast spoiled gradient echo (fast SPGR) sequence, using cardiovascular phase
array coils. A number of 10-sections were acquired with a slice thickness of 8 mm.
Three acquisitions (TR 1.9) were performed at 1, 2, and 4 NEX.

In Figure 6.6, phantom images at 1 NEX before (Figure 6.6a) and after
(Figure 6.6c) filtering are reported, together with the 4 NEX image (Figure 6.6b).

= 9.2 and TE =

(a) (b) (c)

FIGURE 6.6 Phantom images: (a) 1 NEX image, (b) 4 NEX image, and (c) anisotropic-
filtered 1 NEX image.