Abstract:Significance Traditional high-resolution microscopy techniques are limited in imaging within confined and narrow spaces,such as the cavities of animals or the inner chambers of precision instruments,due to their bulky and complex systems.Microscopic endoscopy technology allows for high-resolution observations within cavities by inserting miniature probes.Common types of endoscopes include rigid endoscopes composed of purely optical lenses and electronic endoscopes using image sensors,with diameters typically ranging from millimeters to centimeters.To achieve imaging systems with even smaller diameters,researchers have begun to explore the use of fiber bundles or single fibers for miniature endoscopic imaging.However,these systems typically require gradient refractive index lenses or scanning devices,resulting in diameters much larger than the imaging field and encountering issues such as edge aberrations and honeycomb noise.In recent years,ultrathin lensless multimode fiber(MMF)endoscopes have emerged as a new research hotspot,achieving numerous breakthroughs in imaging modes such as real sample imaging and image transmission. MMFs,as a type of multimode linear system,have historically been regarded as unpredictable due to their rich spatiotemporal modes(phase,amplitude,polarization,wavelength,and pulse delay)and the sensitive and complex mode coupling characteristics.With recent advancements in optical wavefront shaping and optical field measurement technologies,significant strides have been achieved in controlling optical fields within MMFs.This progress positions them as promising candidates for a new generation of minimally invasive super-resolution endoscopic imaging tools.In comparison to traditional endoscopes,MMF endoscopy technology presents several notable advantages.Firstly,it fully exploits the spatial multiplexing capability of fibers,resulting in ultra-high mode density.Moreover,its spatial bandwidth product exceeds that of fiber bundle endoscopes by an order of magnitude under identical probe diameters.Secondly,no additional lens system is required at the fiber probe end,reducing probe size and encapsulation requirements substantially.Thirdly,leveraging MMFs as the transmission medium enables the creation of complex three-dimensional light field distributions at the fiber exit end through encoded wavefront modulation techniques and mode calculations.This facilitates three-dimensional scanning imaging of samples,yielding more comprehensive and detailed sample information than traditional methods.Furthermore,MMFs fabricated from inert and biocompatible hydrogel materials can be directly integrated into disposable medical endoscopic systems.Overall,MMF-based endoscopic detection systems have made significant advancements and are poised to complement traditional endoscopic techniques in achieving high-precision detection in confined spaces.Nonetheless,the feasibility and performance enhancement of this technology in medical and industrial detection applications encounter various challenges.Consequently,summarizing existing research to inform the future rational development of this field is deemed important and necessary. Progress We initially introduced the mechanism of measuring the spatiotemporal optical field transmission matrix in MMFs and then delineated the evolution of MMFs in real sample imaging,encompassing fluorescence imaging,reflection imaging,unlabeled specific imaging,and multimode imaging,along with image transmission imaging methods,and their imaging performance was summarized(Table 1).We also discussed the application of deep learning,metamaterials,adaptive beacons,and other strategies in disturbance-resistant imaging with MMFs.Finally,we looked ahead to the future of MMFs for fast perturbation-resistant,high-pixel,and multifunctional imaging. Conclusions and Prospects MMF imaging is one of the representative achievements with significant influence and wide application in the field of scattering medium imaging,playing an increasingly important role in biomedical,material science,industrial testing,and other fields.Studying the theoretical and technical issues of measuring and imaging the spatiotemporal optical field of MMFs is of great significance for improving imaging spatial resolution,suppressing noise,and obtaining multidimensional imaging information.

Advances in Deep Learning Based Fiber Optic Imaging(Invited)

Spatiotemporal Light Field Measurement and Imaging Applications Based on Multimode Fibers(Invited)

Spatially Variant Deblur and Image Enhancement in a Single Multimode Fiber Imaged by Deep Learning

Deep Tissue Focusing Based on Machine Learning

All-fiber high-speed image detection enabled by deep learning

High-Speed All-Fiber Micro-Imaging with Large Depth of Field

High Accuracy Transmission and Recognition of Complex Images Through Multimode Fibers Using Deep Learning

Dynamic Imaging Through Random Perturbed Fibers Via Physics-Informed Learning

Seeing Through Multimode Fibers with Physics-Assisted Deep Learning

Deep Learning Based Fluorescence Microscopy Imaging Technologies and Applications

High Definition Images Transmission Through Single Multimode Fiber Using Deep Learning and Simulation Speckles

Deep Learning Imaging Through Specialty Multi-Mode Fibers.

All-fiber Ultrafast Image Detection Enabled by Deep Learning

Deep Learning Based Computational Imaging: Status, Challenges, and Future

Deep Learning Imaging Through Fully-Flexible Glass-Air Disordered Fiber

Learning-Based Image Transport Through Disordered Optical Fibers with Transverse Anderson Localization

Deep Learning Model for Ultrafast Multifrequency Optical Property Extractions for Spatial Frequency Domain Imaging.

Edge Detection in Single Multimode Fiber Imaging Based on Deep Learning.

Fiber Bundle Imaging Resolution Enhancement Using Deep Learning.

Deep Learning for Photonic Design and Analysis: Principles and Applications

Fiber neural networks for the intelligent optical fiber communications