Information transport at the extreme limits of diffusion

Most diffuse imaging techniques work in the regime around 10 transport mean free path lengths (e.g., 1cm of human tissue). To realise the potential of non-invasive diffuse optical imaging modalities requires a focused effort to develop techniques in regimes beyond 100 transport mean free paths. The work presented shows evidence that there is imaging information at these extreme scattering regimes in the spatiotemporal distribution of diffuse photons, and information can be enhanced by considering key experimental parameters. Furthermore, we show numerical evidence that there is enough information to reconstruct images of absorbing shapes embedded in highly diffuse materials when using machine learning inverse retrieval algorithms. This work explores the use of highly diffuse light to enable imaging through scattering media well beyond currently accepted limits.