Limits...
lop-DWI: A Novel Scheme for Pre-Processing of Diffusion-Weighted Images in the Gradient Direction Domain.

Sepehrband F, Choupan J, Caruyer E, Kurniawan ND, Gal Y, Tieng QM, McMahon KL, Vegh V, Reutens DC, Yang Z - Front Neurol (2015)

Bottom Line: Our pre-processing method incorporates prior knowledge about the acquired diffusion-weighted signal, facilitating noise reduction.Consequently, it enhances local reconstruction of the orientation distribution function used to define fiber tracks in the brain.The level of improvement in signal-to-noise ratio and in the accuracy of local reconstruction of fiber tracks was significantly improved using our method.

View Article: PubMed Central - PubMed

Affiliation: Centre for Advanced Imaging, The University of Queensland , Brisbane, QLD , Australia ; Queensland Brain Institute, The University of Queensland , Brisbane, QLD , Australia.

ABSTRACT
We describe and evaluate a pre-processing method based on a periodic spiral sampling of diffusion-gradient directions for high angular resolution diffusion magnetic resonance imaging. Our pre-processing method incorporates prior knowledge about the acquired diffusion-weighted signal, facilitating noise reduction. Periodic spiral sampling of gradient direction encodings results in an acquired signal in each voxel that is pseudo-periodic with characteristics that allow separation of low-frequency signal from high frequency noise. Consequently, it enhances local reconstruction of the orientation distribution function used to define fiber tracks in the brain. Denoising with periodic spiral sampling was tested using synthetic data and in vivo human brain images. The level of improvement in signal-to-noise ratio and in the accuracy of local reconstruction of fiber tracks was significantly improved using our method.

No MeSH data available.


Related in: MedlinePlus

Mean of angular error between fibers of digital phantom and the ground truth, measured from evenly sampled lop-DWI (Figure 1A) and unevenly sampled lop-DWI (Figure 1B), across different SNRs. (A) is the results from simulation with b-value = 1,000 s/mm2 and (B) with b-value = 3,000 s/mm2. Asterisks show statistically significant difference in mean values (p < 0.05).
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4290594&req=5

Figure 7: Mean of angular error between fibers of digital phantom and the ground truth, measured from evenly sampled lop-DWI (Figure 1A) and unevenly sampled lop-DWI (Figure 1B), across different SNRs. (A) is the results from simulation with b-value = 1,000 s/mm2 and (B) with b-value = 3,000 s/mm2. Asterisks show statistically significant difference in mean values (p < 0.05).

Mentions: Figure 7 shows that evenly sampled lop-QBI data had slightly higher SNR than unevenly sampled lop-QBI data. The differences in the mean angular error between the two techniques were significant only in four datasets (see asterisks in Figure 7). It should be noted that unevenly sampled lop-DWI outperformed DWI in all the above criteria, but the improvement was slightly weaker than evenly sampled lop-DWI. Moreover, results (not shown here) from raw data acquired with evenly sampled data and with evenly sampled spiral data were almost the same but slightly different from that with unevenly sampled spiral data, which confirms the importance of even distribution of sampling to avoid orientation variance in the distribution of noise (26).


lop-DWI: A Novel Scheme for Pre-Processing of Diffusion-Weighted Images in the Gradient Direction Domain.

Sepehrband F, Choupan J, Caruyer E, Kurniawan ND, Gal Y, Tieng QM, McMahon KL, Vegh V, Reutens DC, Yang Z - Front Neurol (2015)

Mean of angular error between fibers of digital phantom and the ground truth, measured from evenly sampled lop-DWI (Figure 1A) and unevenly sampled lop-DWI (Figure 1B), across different SNRs. (A) is the results from simulation with b-value = 1,000 s/mm2 and (B) with b-value = 3,000 s/mm2. Asterisks show statistically significant difference in mean values (p < 0.05).
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4290594&req=5

Figure 7: Mean of angular error between fibers of digital phantom and the ground truth, measured from evenly sampled lop-DWI (Figure 1A) and unevenly sampled lop-DWI (Figure 1B), across different SNRs. (A) is the results from simulation with b-value = 1,000 s/mm2 and (B) with b-value = 3,000 s/mm2. Asterisks show statistically significant difference in mean values (p < 0.05).
Mentions: Figure 7 shows that evenly sampled lop-QBI data had slightly higher SNR than unevenly sampled lop-QBI data. The differences in the mean angular error between the two techniques were significant only in four datasets (see asterisks in Figure 7). It should be noted that unevenly sampled lop-DWI outperformed DWI in all the above criteria, but the improvement was slightly weaker than evenly sampled lop-DWI. Moreover, results (not shown here) from raw data acquired with evenly sampled data and with evenly sampled spiral data were almost the same but slightly different from that with unevenly sampled spiral data, which confirms the importance of even distribution of sampling to avoid orientation variance in the distribution of noise (26).

Bottom Line: Our pre-processing method incorporates prior knowledge about the acquired diffusion-weighted signal, facilitating noise reduction.Consequently, it enhances local reconstruction of the orientation distribution function used to define fiber tracks in the brain.The level of improvement in signal-to-noise ratio and in the accuracy of local reconstruction of fiber tracks was significantly improved using our method.

View Article: PubMed Central - PubMed

Affiliation: Centre for Advanced Imaging, The University of Queensland , Brisbane, QLD , Australia ; Queensland Brain Institute, The University of Queensland , Brisbane, QLD , Australia.

ABSTRACT
We describe and evaluate a pre-processing method based on a periodic spiral sampling of diffusion-gradient directions for high angular resolution diffusion magnetic resonance imaging. Our pre-processing method incorporates prior knowledge about the acquired diffusion-weighted signal, facilitating noise reduction. Periodic spiral sampling of gradient direction encodings results in an acquired signal in each voxel that is pseudo-periodic with characteristics that allow separation of low-frequency signal from high frequency noise. Consequently, it enhances local reconstruction of the orientation distribution function used to define fiber tracks in the brain. Denoising with periodic spiral sampling was tested using synthetic data and in vivo human brain images. The level of improvement in signal-to-noise ratio and in the accuracy of local reconstruction of fiber tracks was significantly improved using our method.

No MeSH data available.


Related in: MedlinePlus