AI model enables more than a million-fold acceleration of diffuse optical tomography for real-time diagnosis

Article excerpt

Researchers at University of Tsukuba have built an AI model that predicts how light travels through biological tissue for diffuse optical tomography, a noninvasive imaging method used to spot hemorrhages and tumors. The model completes in roughly 2 milliseconds, more than a million times faster than conventional simulation methods. This speed boost could enable real-time diagnosis, transforming how doctors detect abnormalities during clinical procedures.