A Dictionary Learning Method with Total Generalized Variation for MRI Reconstruction

Figure 7

Reconstructed water phantom images at 35% undersampling. (a) The fully sampled image. ((b), (c), (d)) The reconstruction images corresponding to DLMRI, SHTGV, and DLTGV. (e) Mask data with 35% Cartesian sampling. ((f), (g), (h)) Reconstruction error magnitudes for DLMRI, SHTGV, and DLTGV. ((i), (j)) Enlargements of (a), (b), (c), and (d).

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

(i)

(j)