AIR™ Recon DL, our pioneering reconstruction algorithm, challenges the inherent trade-off between SNR, scan time and image resolution. It is not a filter or a post-processing technique.
It improves image quality at the foundational level, and with trained neural networks embedded directly in the reconstruction pipeline, noise and ringing artifacts are removed in the raw data to deliver final reconstructed images with higher SNR and sharpness.
GE Healthcare was the first one introducing a Deep Learning-based reconstruction algorithm working across all anatomies, providing higher quality MR imaging.
Today, more than 600 MR systems were boosted, more than 10 journal articles were published, and more than 1 million patients were scanned with AIR™ Recon DL.
AIR™ Recon DL is probably one of the most impactful innovations over the past decades in MR. After the introduction of new products and software, we always gather words from our early users. They were unanimous in their feedback, saying that AIR™ Recon DL provides sharper and more detailed images with less noise, enables shorter scan times, without any trade off in image quality, reduces number of repeat series, provides greater consistency between patients and technologists, and improves diagnostic confidence. In summary from a radiologist perspective the images are faster to obtain and easier to read.