AIIR stands out as today's most advanced CT image reconstruction architecture, setting a new standard in CT imaging with its unique design. By optimizing noise reduction, enhancing low-contrast detectability, increasing spatial resolution, and reducing artifacts, AIIR ultimately delivers exceptional image quality and superior diagnostic capabilities.

Up to 267% ※

LCD improvement

Up to 150% ※

Spatial resolution improvement

Up to 90% ※

Dose reduction

Up to 98% ※

Image noise reduction

Combining MBIR with deep learning leverages each method's strengths while mitigating their limitations. In AIIR, the data-fidelity term integrates system optics, detector response, and quantum noise models for each scan, preserving detailed anatomical and pathological information from raw projections. While MBIR's high regularization strength can lead to unnatural appearances—especially at low doses—AIIR replaces it with a CNN(Convolutional neural network)-based model capturing complex clinical features through millions of parameters. Together, the two technologies complement each other, allowing more precise noise differentiation and delivers improved image quality.

The 0.5 mm acquisition across all FOVs and collimations unveils finer details with precision. Its fully integrated design, powered by Through-Silicon-Via (TSV) technology, revolutionises detector architecture by reducing the signal conduction path from centimeters to micrometers, significantly lowering electronic noise and delivering ultra-low noise signal output. The 3D anti-scatter grid, precisely aligned with the X-ray source, effectively blocks scattered photons, achieving a scatter-to-primary ratio of less than 8.5%.