Large Field-Of-View Imaging Through Scattering Layers With Optimized Illumination and Localization–Grayscale Fusion

Non-negative matrix factorization (NMF) provides an effective solution for large FOV non-invasive imaging through scattering layers. However, the emerging NMF requires extensive measurement data across multiple encoding patterns.

Furthermore, NMF reconstructions often suffer from loss of grayscale accuracy and the inclusion of background noise. Here, an innovative method is presented that leverages encoding-sparsity optimization (ESO) to decrease the amount of data required by approximately an order of magnitude. Additionally, a precise reconstruction algorithm is introduced using Localization and Grayscale-Fusion (LG-Fusion), which eliminates background noise and extends the FOV to 4.3 times the memory effect range (MER).