super moon

Publish Post

ISMRM 2011

Easy way to fill the hole on the image

Is there are a better way to fill the hole in the image? The function just like the imfill in the matlab. Yes, there is the answer.

The just we need is to find a marker image, which labeled the background and the holes in different value, then find the holes position set the value with object; To label the image, 1. get the implement of the image; 2. apply the region grow find the connected object, define the object which connect with edge of the image as background, label a value to express the background, others labeled a value express the hole; 3. find the hole position and take place the value with the object value.

It’s obvious easy to understand than the morphology method to fill hole.

A Yoga dog

MRI take home point (11) Done

• Most artifacts in PE direction (such as motion) but chemical shift artifact which in FE direction.
• Ringing artifacts by insufficient sample points, Wraparound artifacts by insufficient FOV.
• Gd-chelates’ primary effect is marked shortening of T1; thus heavily T1-weighted sequences are essential for high sensitivity to cancer.
• Gd-chelate contrast agents have the potential to cause a new, potentially serious disease (nephrogenic systemic fibrosis or NSF) in patients who have moderate to end-stage kidney disease at the time of administration of contrast agent

MRI take home point (10)

• Image SNR increases linearly with pixel size.
• When two similarly acquired images are added or subtracted, the resulting image has a noise level that is square root of 2 higher than that in either image.
• The larger the lesion area or contrast, and the lower the noise, the easier the lesion is to detect.
• The rose model describes the relationship between lesion area, contrast, and noise for lesion detection; the larger the lesion area or contrast, and the lower the noise, the easier the lesion is to detect.
• Artifacts can be caused by imperfection of hardware including B0, B1, gradients, shielding, RF coils.
• Cross-talking in multi-slice MRI. Partial-volume artifacts.

MRI take home point (9)

• “Fast STIR” imaging collects multiple phase-encoding views (or lines in k-space) for each
• 180-TI-90-TE/2-180-TE/2-ADC pulse train, making it a practical alternative
• to FSE for obtaining fat-suppressed T2-weighted images.
• Noise is best measured as the standard deviation in the background of an MRI.
• Image SNR increases approximately linearly with static magnetic field strength above 0.5 tesla.