Canon Medical Systems Korea said that it has launched "Advanced intelligent Clear-IQ Engine" (AiCE), the world's first deep-learning reconstruction software, which quickly produces exceptionally detailed computed tomography images with low-noise properties.

Equipped on Canon Medical's top CT line, the Aquilion ONE Genesis, the software uses image reconstruction technology, based on deep convolutional neural network (DCNN), and is capable of delivering CT images with 20 percent lower noise properties and three to four times higher resolution than the existing products.

"The Aquilion ONE Genesis incorporates a forward projection-based iterative reconstruction technology that operates on a model-based iterative reconstruction algorithm and can reconstruct images in three minutes without additional shots," the company said. "AiCE also dramatically removes noise from the video signal and reconstructs the high-quality image quickly."

The noise control scheme of AiCE is superior to the conventional filtered-back projection or hybrid iterative reconstruction, the company added.