Compressive microendoscopy in the presence of noise and misalignment: the translational correlation advantage of graded-index fibers
Multimode fibers (MMFs) offer a minimally invasive platform for deep-tissue and multisite in vivo optical sensing. Computational microendoscopy methods extend them into the high-resolution imaging domain, but are quite sensitive to experimental errors. This study investigates the robustness of microendoscopy based on compressive sensing (CS) by comparing step-index (SI) MMFs with graded-index (GRIN) fibers. Mode-expansion-based simulations are used to assess the accuracy of image reconstruction in the presence of fundamental photon noise and lateral fiber shifts. We show that both fiber types exhibit comparable noise sensitivity, but GRIN fibers are significantly more robust to lateral misalignment errors. These results establish GRIN MMFs as an alignment-tolerant platform for robust computational microendoscopy.