Three-dimensional endoscopic OCT using sparse sampling with a miniature magnetic-driven scanning probe

We propose to apply sparse sampling and compressive sensing (CS) reconstruction in three-dimensional (3D) endoscopic optical coherence tomography (OCT) to reduce the amount of data required in the imaging process. We used a homemade miniature side-imaging magnetic-driven scanning probe with an outer diameter of 1.4 mm in a 1310 nm swept-source OCT system to acquire two-dimensional (2D) circumferential cross-sectional images of an ex vivo pigeon trachea sample. 3D imaging is then achieved by reconstruction from the multiple 2D images acquired while pulling the sample with a translation stage. Given a total translation distance, we achieved sparse sampling by randomizing the step sizes of the translation stage such that the total number of the acquired 2D frames was reduced compared with conventional 3D imaging with equally spaced step positions. We tested the CS reconstruction with reduced 2D frame numbers of 40%, 60%, and 80% compared with the case of equally spaced step positions. The results show that it is possible to recover reasonable OCT images using sparse sampling with CS reconstruction. Compared with the conventional equally spaced sampling method, our method provides a novel way for image acquisition and reconstruction that could significantly reduce the amount of 3D OCT imaging data, and thus the acquisition time.