Fig. 9 | Scientific Reports

Fig. 9

From: Performance of a GPU- and time-efficient pseudo-3D network for magnetic resonance image super-resolution and motion artifact reduction

Fig. 9

Super resolution reconstruction on MR-ART dataset with qualitative comparison to tricubic interpolation. Tricubic interpolation results exhibit pronounced blurring and loss of fine anatomical details, with white–gray matter boundaries appearing poorly defined. In contrast, the SRR images produced by the proposed method show a substantial improvement in image sharpness, recovering fine structural details and yielding markedly sharper and more distinct tissue boundaries. These results highlight the advantage of learning-based super-resolution over conventional interpolation, even under out-of-distribution conditions.

Back to article page