Deep Learning Reconstruction for MRI

Deep Learning technology helps deliver clear, sharp and distinct images. Acquiring high-resolution images traditionally required long scan time. However, with PIQE, high-resolution images are obtained with the same scan time as conventional scans. For those reasons, under the same conditions, PIQE can capture the structure more sharply than AICE with keeping SNR.

Benefit

Common Filter
Because the filter has a rather strong denoise effect, low-resolution images are highly blurred when noise is removed.

PIQE
PIQE reconstruction matrix size768x768 allows for The higher resolution captures the structure more sharply.

PIQE | Brain

This is another example of a high-resolution image of the hippocampus by PIQE. The original image is 256 matrix with low resolution and high SNR.
The scan time is reasonable and the slice thickness is as thin as 2 mm, despite that the image is clear and good contrast.

PIQE | Brain

Using a regular head and neck coil, with a collection matrix 256x256 setting, imaging can be performed in approximately 1 minute. Structures around the pituitary gland are clearly visible.

PIQE | Orbit

Orbit image. Customers does not need to use a dedicated coil such as a circle coil to image the orbit. The image can be taken in only 30 seconds using a normal head/neck coil with a small FOV and low resolution. The structures around the lens are clearly visible.

PIQE | Knee

Until now, it has taken a long time to obtain higher resolution images in the MSK region; with PIQE, higher resolution images can be obtained in a realistic amount of imaging time. PIQE shows the fine structures around cartilage and meniscus.

PIQE | Uterus

PIQE capturing the structure more finely at three times the resolution.