Motion artifact removal and signal enhancement to achieve in vivo dynamic full field OCT

Abstract

We present a filtering procedure based on singular value decomposition to remove artifacts arising from sample motion during dynamic full field OCT acquisitions. The presented method succeeded in removing artifacts created by environmental noise from data acquired in a clinical setting, including in vivo data. Moreover, we report on a new method based on using the cumulative sum to compute dynamic images from raw signals, leading to a higher signal to noise ratio, and thus enabling dynamic imaging deeper in tissues. * https://www.jscholler.com

Dynamic full-field optical coherence tomography: 3D live-imaging of retinal organoids

“…We observed that two successive acquisitions could lead to different perceptual colour maps. We found that the lowest frequencies were slightly unstable (either due to sensor or mechanical instabilities as described in 31 ). We removed this artefact by removing 3% of the lowest frequencies in the PSD.…”

Label-free Functional and Structural Imaging of Liver Microvascular Complex in Mice by Jones Matrix Optical Coherence Tomography

“…However, these methods do not have sensitivity to the tissue metabolism. Recently, OCT-based dynamics imaging techniques have emerged that enable label-free, noninvasive depth-resolved investigation of tissue activity [25][26][27][28][29][30][31] . These dynamics imaging methods use several different techniques, including timefrequency analysis of the OCT signal 25,29,30 , the temporal variance of the time-sequence OCT signal 28 , and the correlation decay speed of the OCT signal 28,31 , and they then visualize the intra-cellular motion on a pixel-by-pixel basis.…”

Retinal axial motion analysis and implications for real-time correction in human retinal imaging

“…In the context of applying dynamic (D-) FFOCT in vivo, the use of singular value decomposition (SVD) processing was recently proposed to filter out the axial displacement of the sample from the local fluctuations linked to intracellular motility [22]. This study showed that axial motion suppression works as long as residual axial motion after correction is smaller than the temporal coherence gate width [22,23]. In FFOCT high-resolution in vivo retinal imaging, a temporal coherence gate of 8 µm is used [10].…”