In this manuscript, a two-dimensional (2D) micro-electro-mechanical system (MEMS)-based, high-speed beam-shifting spectral domain optical coherence tomography (MHB-SDOCT) is proposed for speckle noise reduction and absolute flow rate measurement. By combining a zigzag scanning protocol, the frame rates of 45.2 Hz for speckle reduction and 25.6 Hz for flow rate measurement are achieved for in-vivo tissue imaging. Phantom experimental results have shown that by setting the incident beam angle to ϕ = 4.76° (between optical axis of objective lens and beam axis) and rotating the beam about the optical axis in 17 discrete angular positions, 91% of speckle noise in the structural images can be reduced. Furthermore, a precision of 0.0032 µl/s is achieved for flow rate measurement with the same beam angle, using three discrete angular positions around the optical axis. In-vivo experiments on human skin and chicken embryo were also implemented to further verify the performance of speckle noise reduction and flow rate measurement of MHB-SDOCT.
© 2019 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
Since the introduction of optical coherence tomography (OCT) in 1990s , speckle modulation in OCT structural images and absolute velocity or flow rate measurement have been primary challenges. In clinical applications, speckle noise could suppress the detailed structural information, potentially causing data loss for disease diagnoses. Moreover, the estimation of Doppler angle poses a challenge for flow rate measurements in practical clinical scenarios.
In OCT, speckle noise is inherent due in most part to the random coherence phase shifts caused by different scattering paths of photons within biological tissue. Several hardware [2–11] and software [12–18] based methods have been proposed to reduce speckle noise. Speckle noise can be reduced by averaging the structural images with different speckle patterns which are generated by a variety of hardware perturbation techniques, such as compounding of optical phase [2,3] incident angle [4–8], polarization , spectral band , strain , and volume [12,13]. The software methods have utilized various filters including wavelets , Wiener filters , Bayesian estimation , rotating Kernel transformation  and interval type II fuzzy set  to suppress speckle noise with different distributions compared to sample intensities. Moreover, filter-based speckle reduction algorithms cannot provide any new sample information . In Bo's work, an extended axial focus was achieved by modulating incident angles in a high transverse resolution OCT system . Each of these hardware and software-based approaches has different trade-off between system complexity and data processing complexity.
In addition to structural images, a number of functional OCT schemes have been proposed for disease diagnosis in clinical application, such as Doppler OCT (DOCT) [20–23]. Based on the fact that phase shift can be caused by the movement of red blood cells along optical axis, the axial component of blood flow velocity can then be measured by calculating the phase shift within the time interval of two A-scans. However, the estimation of the Doppler angle between the sample beam and blood flow direction makes it difficult to measure the absolute velocity or flow rate of blood flow. Several approaches have been proposed to solve this problem, such as using multiple incident angles [24–28], 3D vessel geometry , and velocity integration  methods. A triple-beam OCT method that can simultaneously measure the velocities of three incident angles for absolute velocity calculation has been proposed , although the hardware cost is increased. Multiple incident angles can also be achieved by path-length encoded  or time encoded  methods, but they potentially suffer image degradation at large depth and/or increased system complexity. The dual-angle method is only appropriate for the situation where the optical beam is perpendicular to vessels . For the delay-encoded method, a glass plate is needed in half of the sample beam to achieve different incident angles , making it difficult for 3D scanning. The Doppler angle can also be estimated based on the vessel's physical 3D structure . However, the precision of this method could be affected by the vessel size and distribution. In special cases, such as the optical nerve head region of human retina, the blood flow direction can be nearly parallel to sample beam, thereby enabling flow rate measurement by simply integrating the Doppler velocities within the cross-sectional area of vessels .
Recently, we proposed a beam-shifting spectral domain OCT (BS-SDOCT) for speckle reduction and blood flow rate measurement . In this method, the sample beam is shifted in parallel around the optical axis of the objective lens to achieve incident angle modulation. However, the beam-shifting speed is limited by a pair of electrical motorized stages, resulting in bulk motion noise for speckle reduction and non-real-time flow rate measurement. The velocity plots of several heartbeat cycles with different incident angles are measured and matched for flow rate calculation. The beam-shifting angle, which determines the flow rate measurement precision and speckle reduction efficiency, is limited by low beam-shifting speed.
To solve this problem, a 2D micro-electro-mechanical system (MEMS) based high-speed beam-shifting SDOCT (MHB-SDOCT) is proposed in this work to significantly improve the beam-shifting speed. The beam-shifting angle can also be increased to improve the performance of MHB-SDOCT. The zigzag scanning protocol  is used to decrease the time interval between two adjacent sub-steps for bulk motion noise suppression. The frame rates of 45.2 Hz for speckle reduction and 25.6 Hz for flow rate measurement are achieved for real-time measurement. The performance of the proposed technique is tested and verified by both phantom and in-vivo experiments.
2.1 Beam-shifting method
Figure 1(a) shows the schematics of the incidence angle modulation achieved by a 2D MEMS mirror. As the MEMS mirror shifts the sample beam by a distance R on the galvo mirrors in parallel, the incidence angle on the sample is varied by ϕ, which can be calculated using Eq. (1) in Ref . The left panel in Fig. 1(b) shows the path that the sample beam moves step by step on the objective lens; multiple structural images can be acquired during beam-shifting. Each image has a different speckle pattern due to the variation of incident angles. Specklenoise can be reduced by averaging these images. For real-time flow rate measurement, only three incident angles (θ = −2/3π, 0 and 2/3π) instead of five in our previous work  are performed to decrease data acquisition time, as shown in the right panel of Fig. 1 (b). According to Fig. 1(c), the measured three velocities can be expressed byFig. 1(b), ϕ is the shifted incident angle on the sample determined by the beam-shifting distance R shown in Fig. 1(a), , and are the measured mean velocities which can be obtained from cross-sectional Doppler velocity images and structural images. By solving the equation set, the three orthogonal mean velocity components can be obtained and expressed by
2.2 MHB-SDOCT setup and scanning protocol
Figure 2 shows the schematic of the proposed MHB-SDOCT, where a superluminescent diode (SLD) with a center wavelength of 1310 nm and a bandwidth of 60 nm achieves an axial resolution of 12.6 µm in air. An interferometer based on a 2 × 2 fiber coupler with split ratio of 90:10 is used. The optical power from sample arm is approximately 4.1 mW. Lens 1 and 2 have a focal length of 19 mm, and lens 3 and 4 in sample arm are a pair of achromatic lens with focal length of 50 mm. The 2D MEMS (Mirrorcle Technologies, Inc., US) has a mirror of 800 µm diameter and a step response of 200 µs. The galvos (GVS012, Thorlabs, US) have two mirrors with optical aperture of 10 mm diameter and a small angle scanning bandwidth of 1 kHz. The objective lens (L5) is an achromatic lens with focus of 30 mm, giving a lateral resolution of 10 µm. The beam size on L5 has a diameter of 5 mm. The NIR spectrometer (P&P Optica Inc., Waterloo, Canada) has a spectral resolution of 0.365 nm, which is based on a grating with frequency of 892 lines/mm. For signal detection, a line CCD-based camera (Goodrich SU-LDH2) with 1024 pixels and 12 bit depth was used. This camera was operated at its maximum A-scan rate of 91.9 kHz in all of the experiments of this work.
A zigzag scanning protocol, similar to that found in  was used to decrease the time interval between two consecutive sub-steps for bulk motion noise suppression. As shown in Fig. 3, a full B-scan is divided into five sub-steps. Each sub-step consists of 100 A-scans and has a scanning duty cycle of 83%. Each local position is scanned 17 times for speckle reduction (see Fig. 3(a)), achieving effective frame rate of 45.2 Hz.
For flow rate measurement (Fig. 3(b)), three incident angles were used, and 10 consecutive sub-steps were acquired for each incident angle. Because the smallest measureable velocity for inter-line mode with A-scan rate of 91.9 kHz is too large for in-vivo flow rate measurement, inter-frame Doppler phase shift was calculated to improve Doppler velocity sensitivity. The Kasai window  of 9 (frame direction) × 3 (depth direction) pixels was performed to suppress phase noise. Written informed consent was obtained from the volunteer and this study was regulated by the institutional Research Ethics Board.
3.1 Results of speckle reduction
Phantom experiments were performed to quantitatively analyze the performance of our technique, and the speckle reduction results are shown in Fig. 4. The optical phantom is made of agar and intralipid at 20% with fat particle diameter of approximately 0.5 µm. Figure 4(a)-(e) are five representative structural images with different θ. Figure 4(f) shows the averaged structural image of 17 different incident angle images. We averaged the same number of the structural images with same incident angle of ϕ = 0 for comparison, as shown in Fig. 4(g),where it is observed that the plots of pixel count versus intensities for different N (e.g., the number of images averaged) are nearly identical. By performing the proposed technique, new plots with different N have been generated and are shown in Fig. 4(h). It is shown that the pixel intensity distribution varies from Rayleigh distribution towards Gaussian like distribution with increasing N, demonstrating the reduction of speckle noise. Note that the angle interval (Δθ) between two consecutive beam-shifting positions for each N is evenly divided for a fair comparison, Δθ = 360°/(N-1). Figure 4(i) shows the plots of standard deviation (STD) of pixel intensities versus averaging number. It is seen that the STD plots with ϕ = 4.76° is very close to the theoretical model (1/) when N≤17. However, the STD value with N = 33 does not decrease as much as predicted compared to the value with N = 17, presumably because the beam spots at two consecutive beam-shifting positions are mostly overlapped, resulting in no new speckle patterns. Based on the trade-off between averaging number and data acquiring time, 17 is set to be the averaging number for the following speckle reduction experiments. Compared to theoretical model , the achieved speckle noise reduction performance by averaging 17 structural images is approximately 91% which is calculated by dividing the STD shift between red and blue curves by that between black and blue curve.
We performed a second phantom experiment to further verify the performance of the proposed technique. Figure 5(a)-(c) are the structural images from a two-layer phantom, where the first layer is agar with intralipid and the second layer is agar with polymer beads (Polysciences, Inc., US) with a diameter of 10 µm. In the magnified local images, it is found that the boundary (marked by red arrows) between two layers in Fig. 5(b) is much clearly defined. The intensity plots at the marked position by the dashed red line are also shown in Fig. 5(c) for quantitative comparison; it is seen that speckle noise is greatly reduced by MHB-SDOCT. The signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) are calculated using Eq. (6) and Eq. (7) in  respectively from the local regions marked by dashed green rectangles. The histograms are shown in Fig. 5(d) and Fig. 5(e), the improvement of 17.6 dB on SNR and 6.0 on SNR are achieved by MHB-SDOCT.
The two layers of phantom are also mixed to mimic skin lesions due to the different optical properties. The images obtained by standard OCT and MHB-SDOCT are shown in Fig. 5(f) and (g), respectively. In comparison of the two magnified local regions, it is seen that the structural information is better shown in Fig. 5(g), such as those positions marked by red arrows.
The performance of MHB-SDOCT is also verified by human skin images. In this experiment, a local region on the fourth finger of a 29 year-old male volunteer's left hand was scanned. Figure 6(b)-(c) and Fig. 6(f)-(g) are the cross-sectional and en face structural images respectively. By comparing these images, it is found that more structural information is detected by MHB-SDOCT, such as the magnified local regions. SNR and CNR are also calculated for comparison using the intensities within the regions of dashed green rectangles, the improvement of 11.2 dB on SNR and 3.2 on CNR are achieved.
3.2 Results of flow rate measurement
We implemented a flow phantom to test the performance of MHB-SDOCT, where intralipid solution at 0.5% was pumped through a plastic tube with an inner diameter of 0.28 mm by a syringe pump (Harvard Apparatus, Holliston, MA). In both flow phantom and in-vivo experiments, only one sub-step is used, with 10 consecutive sub-steps being acquired for each incident angle, achieving a frequency of 25.6 Hz for flow rate measurement. To best analyze the performance, different flow directions and different distributions of three incident angles were performed for measurement using MHB-SDOCT. The results are shown in Fig. 7. Figure 7(a)-(b) show the measured flow rate at different flow directions indicated by ϕD between optical axis of objective lens and flow direction. The flow rate was set at 0.078 µl/s through the syringe pump. Figure 7(a) shows the cross-sectional Doppler velocity images and the calculated flow rate plot is shown in Fig. 7(b). It can be seen that the averaged flow rate STD of all flow directions is 0.0019 µl/s, and the STD of all mean flow rates with different Doppler angles is 0.0048 µl/s. In flow rate measurement experiments, phase unwrapping technique from  was performed.
By varying the starting angle θ0 of three different incident angles on the objective lens, the calculated results are shown in Fig. 7(c) (cross-sectional images) and Fig. 7(d) (flow rate plot). Here, the flow rate controlled by syringe pump was 0.22 µl/s. It can be seen that the averaged flow rate STD of all starting angles θ0 is 0.0032 µl/s, and the STD of all mean flow rates at different starting angles was 0.0096 µl/s.
In-vivo experiments on 5-day old chicken embryo (Fig. 8(a)) were performed to further verify the performance. Figure 8(b)-(c) are the cross-sectional structural image and the binary mask for velocity calculation, which is obtained by thresholding the structural images. Doppler phase shift is calculated between adjacent sub-steps, and a Kasai window of 9 (frame direction) × 3 (depth direction) pixels is performed to suppress the phase noise. The measured mean velocity plots with three different incident angles are shown in Fig. 8(d). Figure 8(e) shows the calculated flow rate plot. The representative cross-sectional Doppler velocity images at the time points marked by the red arrows in Fig. 8(e) are shown in Fig. 8(f).
4. Discussion and conclusion
In our previous work , a pair of electrical motorized stages was used to investigate the fundamental feasibility of beam-shifting technique. However, due to the limited beam-shifting speed, image registration is needed to compensate the bulk motions for speckle reduction. This also means that velocity calculation must be performed using several heartbeat cycles for each incident angle, and all obtained plots must be registered with each other. All of these shortcomings are addressed by employing a MEMS device. By using the zigzag scanning protocol, the time interval between two consecutive sub-steps is further decreased for bulk motion noise suppression. The frame rates of 45.2 Hz for speckle reduction and 25.6 Hz for flow rate measurement are achieved. The beam-shifting angle is also increased to improve the speckle reduction efficiency and flow rate measurement accuracy. For the beam-shifting angle of 4.76°, 91.2% of speckle could be suppressed by averaging 17 structural images, and a precision of 0.0032 µl/s is achieved for absolute flow rate measurement. Because inter-frame phase shift is calculated for velocity measurement in this work, the smallest measureable velocity is improved compared to . However, the largest flow rate is decreased as a trade-off. If we assume that the blood flow direction is perpendicular to optical axis, the highest flow rate measurable by the current system could reach 0.49 µl/s for the vessels with a diameter of 300 µm.
The main advantage of the proposed technique is that the functions of speckle reduction and flow rate measurement could be performed on the same system setup, and the beam-shifting scheme can be implemented on a standard one-beam SDOCT. Compared to other angular compounding based speckle reduction systems such as Desjardins’s work [4,5] and Bashkansky’s work , MHB-SDOCT is more flexible to the averaged frame numbers and incident angles. Even though optical phase compounding methods for speckle noise reduction were able to approximate theoretical models (such as in Liba's work  or Zhang's work ), blood flow rate was not obtainable from these systems due to fixed incident angle.
Compared to the flow rate measurement systems such as dual-angle [27,28] or velocity integration  based, our proposed method is more robust as there is no restriction for beam scanning and vessel distribution. Previous works have used triple incident angle methods as well, but all have significant draw-backs . utilized a triple beam setup, but the necessity for more interferometers and detection electronics lead to a significantly larger cost compared to the method proposed in this study . was limited by depth-dependent image degradation. In general, increased system complexity (e.g., ) or cost is observed in such schemes. Furthermore, the three incident angles are fixed for all the methods mentioned above. In contrast, MHB-SDOCT demonstrates that both the incident angle and starting angle (θ0) can be flexible, which is useful in the scenarios such as the avoidance of surface reflection.
The MEMS based beam-shifting scheme can also be used for human retinal flow rate measurement based on the fact that human heart beat frequency is approximately 1.00 to 1.67 Hz. In the zigzag scanning protocol, the maximum scanning range and repeat frequency of sub-steps are determined by the fast scanning galvo. In this work, the fast scanning galvo runs at a scanning range of approximately 1 mm (sub-step) and at a frequency of 769 Hz. A pair of galvos can also be used for 2D beam-shifting, although this would require an additional lens pair used in the system, thereby increasing system complexity and cost. It is important to note that the MEMS device used in this work costs less than a pair of galvos.
The beam-shifting angle in MHB-SDOCT is determined by objective lens aperture and beam diameter on objective lens, which can be larger than what used here (4.76°). However, because the structural images with different incident angles on sample are averaged for speckle reduction, MHB-SDOCT suffers the trade-off between speckle reduction performance and lateral resolution degradation. In this work, a lateral resolution degradation of a factor of 3 at the entrance of the Rayleigh range was deemed acceptable as the worst case for beam-shifting angle calculation. Because the resolution degradation plot for beam-shifting angle of 4.76° is not shown in our previous work , we calculated the plot also using Eq. (8) in . The result is shown in Fig. 9, where skin refractive index of 1.35 was used for calculation. Moreover, the optical path is also affected by the distribution of skin surface, resulting in an unpredictable change of lateral resolution of MHB-SDOCT. Depth dependent image degradation in triple beam setup has been observed before, such as in .
In conclusion, we proposed a 2D MEMS based high-speed beam-shifting spectral domain optical coherence tomography (MHB-SDOCT) for in-vivo speckle noise reduction and absolute flow rate measurement. The frame rates of 45.2 Hz for speckle reduction and 25.6 Hz for flow rate measurement were achieved by using a zigzag scanning protocol. The performance of MHB-SDOCT has been verified by both phantom and in-vivo experiments. MHB-SDOCT can be performed on a standard one-beam SDOCT or phase-stable swept source OCT. This work provided a solution for applications where either or both speckle reduction and flow rate measurements are needed.
China Scholarship Council and Natural Sciences and Engineering Research Council of Canada.
This research work is supported by China Scholarship Council, and Canada Research Chair program of Natural Sciences and Engineering Research Council of Canada (NSERC).
The authors declare that there are no conflicts of interest related to this article.
1. D. Huang, E. Swanson, C. Lin, J. Schuman, W. Stinson, W. Chang, M. Hee, T. Flotte, K. Gregory, and C. Puliafito, “Optical coherence tomography,” Science 254(5035), 1178–1181 (1991). [CrossRef] [PubMed]
2. O. Liba, M. D. Lew, E. D. SoRelle, R. Dutta, D. Sen, D. M. Moshfeghi, S. Chu, and A. de la Zerda, “Speckle-modulating optical coherence tomography in living mice and humans,” Nat. Commun. 8, 15845 (2017). [CrossRef] [PubMed]
3. P. Zhang, S. K. Manna, E. B. Miller, Y. Jian, R. K. Meleppat, M. V. Sarunic, E. N. Pugh Jr., and R. J. Zawadzki, “Aperture phase modulation with adaptive optics: a novel approach for speckle reduction and structure extraction in optical coherence tomography,” Biomed. Opt. Express 10(2), 552–570 (2019). [CrossRef] [PubMed]
4. A. E. Desjardins, B. J. Vakoc, G. J. Tearney, and B. E. Bouma, “Speckle reduction in OCT using massively-parallel detection and frequency-domain ranging,” Opt. Express 14(11), 4736–4745 (2006). [CrossRef] [PubMed]
5. A. E. Desjardins, B. J. Vakoc, W. Y. Oh, S. M. R. Motaghiannezam, G. J. Tearney, and B. E. Bouma, “Angle-resolved optical coherence tomography with sequential angular selectivity for speckle reduction,” Opt. Express 15(10), 6200–6209 (2007). [CrossRef] [PubMed]
7. N. Iftimia, B. E. Bouma, and G. J. Tearney, “Speckle reduction in optical coherence tomography by “path length encoded” angular compounding,” J. Biomed. Opt. 8(2), 260–263 (2003). [CrossRef] [PubMed]
8. D. Cui, E. Bo, Y. Luo, X. Liu, X. Wang, S. Chen, X. Yu, S. Chen, P. Shum, and L. Liu, “Multifiber angular compounding optical coherence tomography for speckle reduction,” Opt. Lett. 42(1), 125–128 (2017). [CrossRef] [PubMed]
10. M. Pircher, E. Gotzinger, R. Leitgeb, A. F. Fercher, and C. K. Hitzenberger, “Speckle reduction in optical coherence tomography by frequency compounding,” J. Biomed. Opt. 8(3), 565–569 (2003). [CrossRef] [PubMed]
12. M. R. N. Avanaki, R. Cernat, P. J. Tadrous, T. Tatla, A. G. Podoleanu, and S. A. Hojjatoleslami, “Spatial Compounding Algorithm for Speckle Reduction of Dynamic Focus OCT Images,” IEEE Photonics Technol. Lett. 25(15), 1439–1442 (2013). [CrossRef]
13. M. Szkulmowski, I. Gorczynska, D. Szlag, M. Sylwestrzak, A. Kowalczyk, and M. Wojtkowski, “Efficient reduction of speckle noise in Optical Coherence Tomography,” Opt. Express 20(2), 1337–1359 (2012). [CrossRef] [PubMed]
14. D. C. Adler, T. H. Ko, and J. G. Fujimoto, “Speckle reduction in optical coherence tomography images by use of a spatially adaptive wavelet filter,” Opt. Lett. 29(24), 2878–2880 (2004). [CrossRef] [PubMed]
15. A. Ozcan, A. Bilenca, A. E. Desjardins, B. E. Bouma, and G. J. Tearney, “Speckle reduction in optical coherence tomography images using digital filtering,” J. Opt. Soc. Am. A 24(7), 1901–1910 (2007). [CrossRef] [PubMed]
16. A. Wong, A. Mishra, K. Bizheva, and D. A. Clausi, “General Bayesian estimation for speckle noise reduction in optical coherence tomography retinal imagery,” Opt. Express 18(8), 8338–8352 (2010). [CrossRef] [PubMed]
17. J. Rogowska and M. E. Brezinski, “Evaluation of the adaptive speckle suppression filter for coronary optical coherence tomography imaging,” IEEE Trans. Med. Imaging 19(12), 1261–1266 (2000). [CrossRef] [PubMed]
18. P. Puvanathasan and K. Bizheva, “Speckle noise reduction algorithm for optical coherence tomography based on interval type II fuzzy set,” Opt. Express 15(24), 15747–15758 (2007). [CrossRef] [PubMed]
19. E. Bo, X. Ge, X. Yu, J. Mo, and L. Liu, “Extending axial focus of optical coherence tomography using parallel multiple aperture synthesis,” Appl. Opt. 57(13), 3556–3560 (2018). [CrossRef] [PubMed]
20. J. A. Izatt, M. D. Kulkarni, S. Yazdanfar, J. K. Barton, and A. J. Welch, “In vivo bidirectional color Doppler flow imaging of picoliter blood volumes using optical coherence tomography,” Opt. Lett. 22(18), 1439–1441 (1997). [CrossRef] [PubMed]
21. S. Yazdanfar, M. Kulkarni, and J. Izatt, “High resolution imaging of in vivo cardiac dynamics using color Doppler optical coherence tomography,” Opt. Express 1(13), 424–431 (1997). [CrossRef] [PubMed]
22. Z. Chen, T. E. Milner, S. Srinivas, X. Wang, A. Malekafzali, M. J. C. van Gemert, and J. S. Nelson, “Noninvasive imaging of in vivo blood flow velocity using optical Doppler tomography,” Opt. Lett. 22(14), 1119–1121 (1997). [CrossRef] [PubMed]
23. Y. Zhao, Z. Chen, C. Saxer, S. Xiang, J. F. de Boer, and J. S. Nelson, “Phase-resolved optical coherence tomography and optical Doppler tomography for imaging blood flow in human skin with fast scanning speed and high velocity sensitivity,” Opt. Lett. 25(2), 114–116 (2000). [CrossRef] [PubMed]
24. R. Haindl, W. Trasischker, A. Wartak, B. Baumann, M. Pircher, and C. K. Hitzenberger, “Total retinal blood flow measurement by three beam Doppler optical coherence tomography,” Biomed. Opt. Express 7(2), 287–301 (2016). [CrossRef] [PubMed]
25. A. Wartak, R. Haindl, W. Trasischker, B. Baumann, M. Pircher, and C. K. Hitzenberger, “Active-passive path-length encoded (APPLE) Doppler OCT,” Biomed. Opt. Express 7(12), 5233–5251 (2016). [CrossRef] [PubMed]
26. A. Wartak, F. Beer, B. Baumann, M. Pircher, and C. K. Hitzenberger, “Adaptable switching schemes for time-encoded multichannel optical coherence tomography,” J. Biomed. Opt. 23(5), 1–12 (2018). [CrossRef] [PubMed]
27. R. M. Werkmeister, N. Dragostinoff, M. Pircher, E. Götzinger, C. K. Hitzenberger, R. A. Leitgeb, and L. Schmetterer, “Bidirectional Doppler Fourier-domain optical coherence tomography for measurement of absolute flow velocities in human retinal vessels,” Opt. Lett. 33(24), 2967–2969 (2008). [CrossRef] [PubMed]
28. C. J. Pedersen, D. Huang, M. A. Shure, and A. M. Rollins, “Measurement of absolute flow velocity vector using dual-angle, delay-encoded Doppler optical coherence tomography,” Opt. Lett. 32(5), 506–508 (2007). [CrossRef] [PubMed]
29. Z. Ma, A. Liu, X. Yin, A. Troyer, K. Thornburg, R. K. Wang, and S. Rugonyi, “Measurement of absolute blood flow velocity in outflow tract of HH18 chicken embryo based on 4D reconstruction using spectral domain optical coherence tomography,” Biomed. Opt. Express 1(3), 798–811 (2010). [CrossRef] [PubMed]
30. B. Baumann, B. Potsaid, M. F. Kraus, J. J. Liu, D. Huang, J. Hornegger, A. E. Cable, J. S. Duker, and J. G. Fujimoto, “Total retinal blood flow measurement with ultrahigh speed swept source/Fourier domain OCT,” Biomed. Opt. Express 2(6), 1539–1552 (2011). [CrossRef] [PubMed]
31. C. Chen, W. Shi, R. Deorajh, N. Nguyen, J. Ramjist, A. Marques, and V. X. D. Yang, “Beam-shifting technique for speckle reduction and flow rate measurement in optical coherence tomography,” Opt. Lett. 43(24), 5921–5924 (2018). [CrossRef] [PubMed]
32. C. Chen, W. Shi, R. Reyes, and V. X. D. Yang, “Buffer-averaging super-continuum source based spectral domain optical coherence tomography for high speed imaging,” Biomed. Opt. Express 9(12), 6529–6544 (2018). [CrossRef]
33. V. Yang, M. Gordon, B. Qi, J. Pekar, S. Lo, E. Seng-Yue, A. Mok, B. Wilson, and I. Vitkin, “High speed, wide velocity dynamic range Doppler optical coherence tomography (Part I): System design, signal processing, and performance,” Opt. Express 11(7), 794–809 (2003). [CrossRef] [PubMed]
34. S. Xia, Y. Huang, S. Peng, Y. Wu, and X. Tan, “Robust phase unwrapping for phase images in Fourier domain Doppler optical coherence tomography,” J. Biomed. Opt. 22(3), 36014 (2017). [CrossRef] [PubMed]