Open Access
Add to BibSonomyAbstract
To improve the lateral resolution of the blood vessels along arbitrary direction out of focus in photoacoustic microscopy (PAM), we propose an adaptive synthetic-aperture focusing technique (ASAFT) for microvasculature imaging which can be automatically applied to each branch of blood vessels, based on our previous two-dimensional (2D) SAFT. The ASAFT is validated both in the phantom study and in vivo imaging. The results demonstrate that ASAFT can provide images of blood vessels with better lateral resolution both at different depths and along various directions compared with one-dimensional and 2D SAFT.
©2012 Optical Society of America
1. Introduction
Vasculature and microcirculation transport the necessities for normal metabolism to all parts of living tissues and organs, and collect the metabolic waste as well [1,2]. The abnormal anatomy of the vasculature often affects the homeostasis of the organisms and ultimately leads to tissue inviability [3]. Many worldwide vicious diseases, such as diabetes [4], arteriosclerosis [4], stroke [5,6], and tumor angiogenesis [7,8], are greatly related to the changes of the structural features of the vascular tree. Other physiological and pathological processes, including the healing of wounds [9,10], vasodilation, and vasomotion [1,11], also induce changes in the vasculature. Therefore, in vivo vasculature imaging with high spatial resolution is extremely important for disease diagnosis, clinical treatment, and the physiological mechanisms study [4,7,12]. In in vivo vasculature and microcirculation studies, intravital microscopy (IVM) is the gold standard [13]. Unfortunately, it lacks depth resolution which is crucial for three-dimensional (3D) vasculature imaging. Current well-established clinical imaging modalities which can be adopted for vascular imaging include computed tomography (CT) [14], magnetic resonance imaging (MRI) [15], positron emission tomography (PET) [16] and ultrasonography (US) [17]. However, they generally lack either sufficient spatial resolution or satisfactory contrast for microcirculation imaging in vivo. Furthermore, CT and PET require the intravenous administration of exogenous contrasts and may consequently cause carcinogensis. Photoacoustic microscopy (PAM) which is developed rapidly in recent years, combines the excellent optical contrast and the high ultrasonic spatial resolution [18, 19]. It has been proven to be an emerging tool for imaging tumor angiogenesis [20, 21], subcutaneous microvasculature [22], oxygenation monitoring in blood vessels [23, 24], and so on. In PAM, a short-pulsed laser beam irradiates the tissue to generate the ultrasonic wave based on the thermoelastic effect. And the photoacoustic wave is detected by the ultrasonic transducer. The rich contrast is mainly determined by the pulse energy absorbed by hemoglobin and other chromophores in tissue [25, 26]. Since the overlapping ultrasonic focus is much smaller than the optical focus in acoustic-resolution PAM, the lateral resolution is mainly determined by the focused ultrasonic transducer.
To achieve high resolution at the focus, a spherically focused transducer with high central frequency and a large numerical aperture is usually adopted to receive the generated photoacoustic waves. However, it has limited depth of focus, which makes the lateral resolution degrade significantly out of focus [18, 27]. To solve this problem, Zhang et al. proposed the idea of auto-fit scan to acquire the 3D subcutaneous microvasculature in a subcutaneous tumor [28]. However, it required more data acquisition for the contour scan. And only part of the vessels can be controlled in focus since the focused transducer had the focal zone with only 200~300 μm in length. Recently, some research groups used synthetic-aperture focusing technique (SAFT) to extend the depth of focus which was widely employed in radar and sonar. Even with coherence and adaptive weighting factor, one-dimensional (1D) SAFT was demonstrated that it could improve the lateral resolution and signal to noise ratio (SNR) along the B-scan direction in the PAM [27, 29, 30]. To improve the lateral resolution and SNR both in the two lateral directions simultaneously, Deng et al. suggested a two-dimensional (2D) SAFT with coherence weighting factor (CF) [31]. However, the lateral resolution out of the focus in two lateral directions was still worse than that at the focus. This is mainly because of a potential assumption that the photoacoustic waves generated by the absorbers have a spherical wave-front. But the blood vessels absorb laser pulse and produce photoacoustic waves with a cylindrical wave-front [32]. Under this condition, the vessels perpendicular to the synthetic direction can be well reconstructed. With the complex distribution of the vascular network in vivo, only parts of the blood vessels at the focus can be imaged with high resolution in practice.
In this study, we present an adaptive synthetic-aperture focusing technique (ASAFT) with CF used for vascular network imaging in vivo to improve the resolution of the vessels with arbitrary direction. Firstly, the orientations of the vessel branches are extracted from the maximum amplitude projection (MAP) processed with the 2D SAFT and virtual-point-detector concept. Secondly, the SAFT is applied along the direction that is perpendicular to the corresponding vessel branch to improve the lateral resolution out of focus. To demonstrate the efficiency of the proposed method, we use a phantom composed of several 25 μm tungsten wires and the results are also quantified. Furthermore, the diameters of the blood vessels are estimated in vivo after processing using different SAFT methods. And the results acquired using stereomicroscope is regarded as criterion.
2. The adaptive synthetic-aperture focusing technique
In acoustic-resolution PAM, the coaxially overlapping optical focus is much wider than the ultrasonic focus. The photoacoustic waves generated by the biological tissue within a certain solid angle are received by the transducer [27]. We regard the geometric focal point of the transducer as a virtual point detector. If a 2D raster scan is performed in a horizontal plane, a 3D volumetric image can be obtained. The depth-independent lateral resolution in the B-scan image can be achieved with 1D SAFT used in the same plane [27, 31, 33]. Also, a 2D SAFT is suitable to yield an image with isotropic resolution simultaneously in the two lateral directions out of focus [31]. However, when the absorbers are vasculature, they will generate photoacoustic waves with a cylindrical wave-front. In this case, only the blood vessels perpendicular to synthetic direction can get the best lateral resolution. To image the blood vessels with high lateral resolution in all the orientations and both in and out of the focus, we establish an ASAFT with CF based on the orientations of the vessel branches and virtual-point-detector concept, as illustrated in Fig. 1 .
In the proposed method, the first step was to determine the directions of different vessel branches. After acquiring the original volumetric data, 2D SAFT was performed and the corresponding MAP was achieved. In this approach, the geometrical structures of the blood vessels can be considered as tubular. To remove the noise and systematic errors in the image, the MAP formed by 2D SAFT was further processed by using multiscale vesselness filtering to enhance equally tubular structures with various diameters. Then, a binary image was acquired by the amplitude threshold [34, 35] and the corresponding image skeleton was extracted by optimization of the bwmorph function using the MATLAB Image Processing Toolbox (R2008a, Mathworks). In ASAFT, each branch of the blood vessels was synthesized individually. For the corresponding skeleton, the direction of a point was determined by the point and its 15 consecutive points around. Here, the first-order linear fitting was adopted to fit these points to determine the direction. Based on the orientations of the vessel branches, the signals which are delayed appropriately relative to the virtual point detector are summed across the cross section of each blood vessel as follows [31]:
where , . is the angle determined by the direction of vessel branch and the B-scan direction. [] describes the nearest integer towards minus infinity and k indicates the unit pixel along the direction where ASAFT was used. is the recorded signal at each position and denotes the time delay from the synthetic focal point to the virtual detector along the synthetic direction. represents the signal processed using the ASAFT. N is the maximum number of scan lines summed and associated with the angular extent of the ultrasonic radiation pattern.To further improve the SNR and spatial resolution, a data-dependent coherence factor (CF) was used as a focusing-quality index for each SAFT imaging point [31]. After the ASAFT and CF were carried out, the improved image of vascular networks were achieved.
3. Results
The details of the PAM used in this investigation can be found in our previous reports [25,31]. The system adopted a transducer with center frequency of 50 MHz to collect the ultrasound. With a spherically focusing lens, a focal length of 6.7 mm and a focal zone of 300 μm in length were achieved. These enabled the system to exhibit lateral and axial resolutions of 45µm and 15 µm at the focus, respectively. The imaging depth is 3 mm at least.
3.1 Phantom study
To evaluate the efficiency of the proposed method, seven 25 μm-tungsten wires fixed with 1% agar were used in the phantom study. The tungsten wires were all in different directions and at different depths to simulate the distribution of vascular networks. In the experiment, the phantom was adjusted far from the focus and then a 2D raster scan was performed on an area of 9.4 × 7 mm2 with a step size of 20 µm. After the original data was acquired, it was interpolated by a factor of 2 in all directions to increase the delay accuracy. Then, different SAFT were performed on the interpolated data. The acoustic velocity of 1.5 mm/µs was used to calculate the time delay. Figure 2(a) displays the projected C-scan images with no further processing. Figure 2(b), 2(c), and 2(d) show the image processed by 1D SAFT in the B-scan direction, 2D SAFT in both lateral directions, and adaptive SAFT, respectively. The 3D distribution of tungsten wires can be visualized from the volumetric imaging of the phantom as shown in Fig. 2(f).
Fig. 2 Images of the phantom containing seven 25 μm-tungsten wires, which were imaged out of focus. (a) the MAP image acquired from the original data; The dotted lines marked by t1 and t2 are the cross-sectional profile of the corresponding tungsten wires used for further quantitative analysis; The tungsten wires indicated with 1-6 are all in different directions and at different depths; (b)-(d) the MAP image after processed with 1D SAFT, 2D SAFT and ASAFT, respectively; (e) two-dimensional skeleton image extracted from (c); (f) the three-dimensional rendering of the phantom.
From Fig. 2(a) and 2(f), it can be seen that parts of tungsten wires indicated by number 2 and 3 are near the focus and correspondingly present high intensity. The tungsten wires labeled as 4-6 show the lateral resolution deteriorates significantly, because they are far from the focus. With 1D SAFT adopted in the B-scan direction, the tungsten wires marked as 2, 3 and 4 were imaged with improved lateral resolution, whereas others were not. Based on 2D SAFT, the lateral resolution of the entire scanning window was improved to become homogeneous. However, some tungsten wires present poorer lateral resolution compared with Fig. 2(b), such as the tungsten wires indicated by 2 and 3. Figure 2(d) represents the ASAFT and CF weighted image and all the tungsten wires exhibit the best lateral resolution.
In order to quantitatively analyze the lateral resolution, the tungsten wires marked with t1 and t2 in Fig. 2 were selected and processed by different SAFT methods. Here, the lateral resolution is determined by the full width at half maximum (FWHM) of the deconvolution of the corresponding cross-sectional profile and tungsten wire diameter. It can be seen that the original lateral resolution at the position marked by t1 and t2 in Fig. 2(a) degrades to 116.3μm and 155.6μm out of the focus. Correspondingly, the lateral resolutions are improved to 99μm and 78.5μm with 1D SAFT, 97.5μm and 88.2μm with 2D SAFT. Using ASAFT, the lateral resolution at the position marked by t1 and t2 are further improved to be 57.7μm and 49.5μm. The results indicate that the resolution at the position indicated by t2 is significant improved using 1D SAFT, whereas it is not at the position indicated by t1. This is because tungsten wire t1 is nearly parallel to the B-scan direction, where 1D SAFT only provide the better lateral resolution along the B-scan direction in the PAM. In 2D SAFT, the delayed signals are summed along both lateral planes simultaneously to yield the images of tungsten wires with consistent improved resolution. With ASAFT, the synthesis was performed along the direction perpendicular to the tungsten wire. Therefore, the best lateral resolution can be achieved, compared to 1D and 2D SAFT. Figure 3 also shows that the proposed method increases the SNR of cross-sectional profiles of t1 and t2 by up to 27.77dB and 29.86dB, respectively, comparing to original values. The results are close to that processed with 2D SAFT, where the corresponding SNR are 27.95dB and 29.8 dB. However, they are much higher than 19.79dB and 21.43dB which are acquired using 1D SAFT. The improved SNR uses 2D SAFT mainly because synthesis is performed in two directions. With our method, similar performance of SNR can be achieved with only one-direction synthesis. Therefore, we can conclude that our method performs better than 1D and 2D SAFT in resolution, and shows similar SNR to 2D SAFT.
3.2 In vivo imaging
To further validate the adaptive SAFT in small animal imaging, the vascular distribution in the dorsal subcutaneous vessels of a BALB/c mouse with the weight of 35 g was used. The mouse was anesthetized with a mixture of 0.2g/kg a-chloralose and 1g/kg urethane administered intraperitoneally. Before imaging, the hair on the region of interest was removed by a commercial human hair removing lotion. Then, an area of 13.2 × 11.7 mm2 was chosen for a C-scan conducted with a step size of 30µm. During the operation, the body temperature of the mouse was kept constant at 37 ± 0.5°C using a heating pad. After about 100-minute C-scan, the mouse was euthanatized with an overdose of anesthetic. A photography of the vascular distribution from the inner surface of the C-scan area was taken by a stereomicroscope (SZX12, Olympus, Japan) equipped with a charge coupled device (CCD) camera (QE 270 XS, PCO Pixelfly, Germany). The resolution of the stereomicroscope is better than 10 µm, considering the numberical aperture of the stereomicroscope and effective resolution of the CCD camera. The animal experiment in the research was approved by the Committee of Experimental Animals of Hubei Province, and the procedures were carried out according to the routine animal-care guidelines.
Figure 4 illustrates the MAP of vascular distribution with parts of vessels which are out of the focus, as well as the results processed with different SAFT methods are also shown. The original MAP image in Fig. 4(a) indicates that only the vessels in the center of the scanning window were imaged with high resolution. Compared with Fig. 5(a) , it is obvious that these vessels were scanned at the focus, including the vessels marked as 1 and 2. Figure 4(a) also shows that the vessels at the margin of the imaging window were blurred. This is because these vessels are out of focus and the lateral resolution degrades, such as vessel 3 and 4. It can be seen from Fig. 5(a) that vessel 4 is about 900μm below the focus.
Fig. 4 Images of the vascular distribution in the dorsal dermis in vivo: (a) original image, (b) 1D SAFT image, (c) 2D SAFT image, (d) ASAFT image and (e) the photography of the vascular network acquired using a stereomicroscope. (f) the 2D vessel skeleton image extracted from (c). The vessels indicated by 1-4 are also shown in Fig. 5(a) and these vessels marked by the dashed boxes are used for further analysis.
Fig. 5 In vivo B-scan images whose position is marked with the dashed line in Fig. 4(a) and processed with different methods: (a) original image, (b) 1D SAFT, (c) 2D SAFT, (d) ASAFT. The vessels labeled with 1 and 2 are at the focus, while 3 and 4 are out of focus.
Table 1 demonstrates the diameter of the vessels labeled with v1, v2 and v3 which are acquired using different methods in Fig. 4. Here, the error is defined as the relative difference between the vessel diameters obtained by PAM and the stereomicroscope. As a standard, Fig. 4(e) represents that the diameters of the vessels, indicated by v1, v2 and v3, are 100, 70 and 130μm, respectively. The original photoacoustic images in Fig. 4(a) shows the corresponding diameters are 300, 240 and 700μm with a large error of 200%, 242% and 438.5%. With 1D SAFT, the diameters of vessel 1 and 3 extracting from Fig. 4(b) were measured with decreased error, while vessel 2 was not. Based on 2D SAFT, all the three vessels were imaged with improved resolution. When ASAFT was applied to the volumetric data, only small difference existed and the diameters of vessel 1, 2 and 3 were 110, 100 and 145μm with a decreased error of 10%, 42.9% and 11.5% respectively, which are close to the standard. Considering the resolution of our system, these results are close to the utmost performance of system limited by the numerical aperture of the ultrasound transducer. This further demonstrates 1D SAFT can only improve the lateral resolution in certain direction and 2D SAFT provides a more homogeneous image. Based on the fact that the blood vessels absorbed the laser light to generate the ultrasound with a cylindrical wave-front, ASAFT adopts the synthesis direction that is perpendicular to the blood vessels. Thus, the vasculature can be imaged with high precision at various depths as well as in different directions. Comparing Fig. 4(c) with Fig. 4(d), it becomes apparent that additional features (red arrow) exist in the latter, which is the real structures as the standard (Fig. 4(e)). This further demonstrates that the proposed method improves not only the spatial resolution but also the SNR.
Table 1. Comparison of the Diameters Measured by Stereomicroscope and PAM
4. Conclusion and discussion
In this study, we presented an ASAFT in conjuction with the CF to improve the lateral resolution out of focus in PAM for microvasculature imaging. Using ASAFT, the depth of the focus was extended. This enables the 3D vascular network both at and out of the focus to be imaged with high resolution. The improved performance by applying our method was validated in the phantom study and in vivo investigation. Compared to 1D and 2D SAFT, the ASAFT can provide the images of blood vessels with better performance at different depths and in different directions.
It should be noted that the proposed method is based on the tubular structures and high contrast images acquired by PAM. The high contrast vascular images ensure that the blood vessels can be identified easily. Consequently, the direction of the tubular structures of blood vessels can be acquired and each branch of the vessels can be synthesized with SAFT along the direction perpendicular to the corresponding blood vessel in ASAFT. Thus, our method can provide quantitatively correct information for vasculature imaging. And due to the high-contrast of the images acquired by PAM, as well as the capability of identifying imaging objects, our method can be extended easily. However, it is quite different when the absorbers are point targets such as the dots in Fig. 2(a)- 2(d). In this case, the ultrasound received by the transducer is considered to have a spherical wave-front. Synthesis performed in more directions simultaneously can provide quantitatively correction information for these pointlike absorbers as we demonstrated before [31]. In our application, most absorbers are vessels which are tubular. Therefore, the ASAFT we proposed is now particularly designed for microvasculature imaging.
From Table 1, we can see that the v2 showed larger error compared to v3 whose distance from focus was more than twice of that of v2. Although the lateral resolution out of focus can be improved to about 55μm in the phantom study with the proposed method, it will deteriorate in in vivo imaging since more high-frequency components of photoacoustic signal will be lost. The diameters of v2 and v3 are 70μm and 130μm, respectively, measured by a stereomicroscope. It can be seen that the diameter of v2 is close to the lateral resolution limit of PAM. The error in our manuscript is defined as the relative difference between the vessel diameter obtained by PAM and the stereomicroscope measurement. Therefore, a larger error exists in the measured diameter of v2 using PAM than that of v3. This also suggests that our PAM is more appropriate for quantitative study of structure of vessels with diameters larger than 100μm.
Even with ASAFT, the lateral resolution out of the focus is still worse than that at the focus in our study. This may be caused by the following reasons: first, the mismatch between the spherically focused transducer used in PAM and ultrasound generated by a tungsten wire or a blood vessel with a cylindrical wave-front can deteriorate the lateral resolution. Second, only the directions of different vessels or tungsten wires in two-dimensional MAP were used, while a three-dimensional distribution existed. Third, the difference of acoustic velocity in water and tissue was not considered in both the phantom study and in vivo imaging. All these reasons might induce errors in time delay calculated in SAFT. They will be addressed in our future work.
Acknowledgments
This work was supported by the National Major Scientific Research Program of China (Grant No. 2011CB910401), Science Fund for Creative Research Group of China (Grant No.61121004) and National Key Technology R&D Program of China (Grant No. 2012BAI23B02).
References and links
1. S. Hu and L. V. Wang, “Photoacoustic imaging and characterization of the microvasculature,” J. Biomed. Opt. 15(1), 011101 (2010). [CrossRef] [PubMed]
2. J.k.-J. Li, Dynamics of the Vascular System (World Scientific, Singapore, 2004).
3. J. P. Culver, T. Durduran, D. Furuya, C. Cheung, J. H. Greenberg, and A. G. Yodh, “Diffuse optical tomography of cerebral blood flow, oxygenation, and metabolism in rat during focal ischemia,” J. Cereb. Blood Flow Metab. 23(8), 911–924 (2003). [CrossRef] [PubMed]
4. A. M. Mendonca and A. Campilho, ““Segmentation of retinal blood vessels by combining the detection of centerlines and morphological reconstruction,” IEEE,” IEEE Trans. Med. Imag. 25(9), 1200–1213 (2006). [CrossRef]
5. A. Livnat, M. Tolmasov, E. B. Michaely, and A. Mayevsky, “Real-time monitoring of mitochondrial function and cerebral blood flow following focal ischemia in rats,” J. Innovative Opt. Health Sci. 1(01), 63–69 (2008). [CrossRef]
6. A. Durukan and T. Tatlisumak, “Acute ischemic stroke: overview of major experimental rodent models, pathophysiology, and therapy of focal cerebral ischemia,” Pharmacol. Biochem. Behav. 87(1), 179–197 (2007). [CrossRef] [PubMed]
7. J. R. Less, T. C. Skalak, E. M. Sevick, and R. K. Jain, “Microvascular architecture in a mammary carcinoma: branching patterns and vessel dimensions,” Cancer Res. 51(1), 265–273 (1991). [PubMed]
8. Y. Q. Lao, D. Xing, S. H. Yang, and L. Z. Xiang, “Noninvasive photoacoustic imaging of the developing vasculature during early tumor growth,” Phys. Med. Biol. 53(15), 4203–4212 (2008). [CrossRef] [PubMed]
9. E. Zcharia, R. Zilka, A. Yaar, O. Yacoby-Zeevi, A. Zetser, S. Metzger, R. Sarid, A. Naggi, B. Casu, N. Ilan, I. Vlodavsky, and R. Abramovitch, “Heparanase accelerates wound angiogenesis and wound healing in mouse and rat models,” FASEB J. 19(2), 211–221 (2005). [CrossRef] [PubMed]
10. A. A. Tandara and T. A. Mustoe, “Oxygen in wound healing--more than a nutrient,” World J. Surg. 28(3), 294–300 (2004). [CrossRef] [PubMed]
11. C. Aalkjaer and H. Nilsson, “Vasomotion: cellular background for the oscillator and for the synchronization of smooth muscle cells,” Br. J. Pharmacol. 144(5), 605–616 (2005). [CrossRef] [PubMed]
12. R. Wild, S. Ramakrishnan, J. Sedgewick, and A. W. Griffioen, “Quantitative assessment of angiogenesis and tumor vessel architecture by computer-assisted digital image analysis: effects of VEGF-toxin conjugate on tumor microvessel density,” Microvasc. Res. 59(3), 368–376 (2000). [CrossRef] [PubMed]
13. S. Hu, K. Maslov, and L. V. Wang, “Noninvasive label-free imaging of microhemodynamics by optical-resolution photoacoustic microscopy,” Opt. Express 17(9), 7688–7693 (2009). [CrossRef] [PubMed]
14. U. Hoffmann, M. Ferencik, R. C. Cury, and A. J. Pena, “Coronary CT angiography,” J. Nucl. Med. 47(5), 797–806 (2006). [PubMed]
15. M. Neeman, “Functional and molecular MR imaging of angiogenesis: seeing the target, seeing it work,” J. Cell. Biochem. Suppl. 87(S39), 11–17 (2002). [CrossRef] [PubMed]
16. J. H. Rudd, K. S. Myers, S. Bansilal, J. Machac, C. A. Pinto, C. Tong, A. Rafique, R. Hargeaves, M. Farkouh, V. Fuster, and Z. A. Fayad, “Atherosclerosis inflammation imaging with 18F-FDG PET: carotid, iliac, and femoral uptake reproducibility, quantification methods, and recommendations,” J. Nucl. Med. 49(6), 871–878 (2008). [CrossRef] [PubMed]
17. U. Schminke, L. Motsch, B. Griewing, M. Gaull, and C. Kessler, “Three-dimensional power-mode ultrasound for quantification of the progression of carotid artery atherosclerosis,” J. Neurol. 247(2), 106–111 (2000). [CrossRef] [PubMed]
18. K. Maslov, H. F. Zhang, and L. V. Wang, “Photoacoustic generation of focused quasi-unipolar pressure pulses,” J Innov Opt Health Sci 3(4), 247–253 (2010). [CrossRef] [PubMed]
19. H. F. Zhang, K. Maslov, G. Stoica, and L. V. Wang, “Functional photoacoustic microscopy for high-resolution and noninvasive in vivo imaging,” Nat. Biotechnol. 24(7), 848–851 (2006). [CrossRef] [PubMed]
20. M. L. Li, J. C. Wang, J. A. Schwartz, K. L. Gill-Sharp, G. Stoica, and L. V. Wang, “In-vivo photoacoustic microscopy of nanoshell extravasation from solid tumor vasculature,” J. Biomed. Opt. 14(1), 010507 (2009). [CrossRef] [PubMed]
21. D. Pan, M. Pramanik, A. Senpan, J. S. Allen, H. Y. Zhang, S. A. Wickline, L. V. Wang, and G. M. Lanza, “Molecular photoacoustic imaging of angiogenesis with integrin-targeted gold nanobeacons,” FASEB J. 25(3), 875–882 (2011). [CrossRef] [PubMed]
22. H. F. Zhang, K. Maslov, M. L. Li, G. Stoica, and L. V. Wang, “In vivo volumetric imaging of subcutaneous microvasculature by photoacoustic microscopy,” Opt. Express 14(20), 9317–9323 (2006). [CrossRef] [PubMed]
23. H. F. Zhang, K. Maslov, M. Sivaramakrishnan, G. Stoica, and L. H. V. Wang, “Imaging of hemoglobin oxygen saturation variations in single vessels in vivo using photoacoustic microscopy,” Appl. Phys. Lett. 90(5), 053901 (2007). [CrossRef]
24. V. Tsytsarev, S. Hu, J. J. Yao, K. Maslov, D. L. Barbour, and L. V. Wang, “Photoacoustic microscopy of microvascular responses to cortical electrical stimulation,” J. Biomed. Opt. 16(7), 076002 (2011). [CrossRef] [PubMed]
25. X. Q. Yang, X. Cai, K. Maslov, L. H. Wang, and Q. M. Luo, “High-resolution photoacoustic microscope for rat brain imaging in vivo,” Chin. Opt. Lett. 8(6), 609–611 (2010). [CrossRef]
26. L. V. Wang, “Multiscale photoacoustic microscopy and computed tomography,” Nat. Photonics 3(9), 503–509 (2009). [CrossRef] [PubMed]
27. M. L. Li, H. F. Zhang, K. Maslov, G. Stoica, and L. V. Wang, “Improved in vivo photoacoustic microscopy based on a virtual-detector concept,” Opt. Lett. 31(4), 474–476 (2006). [CrossRef] [PubMed]
28. H. F. Zhang, K. Maslov, and L. V. Wang, “Automatic algorithm for skin profile detection in photoacoustic microscopy,” J. Biomed. Opt. 14(2), 024050 (2009). [CrossRef] [PubMed]
29. C. K. Liao, M. L. Li, and P. C. Li, “Optoacoustic imaging with synthetic aperture focusing and coherence weighting,” Opt. Lett. 29(21), 2506–2508 (2004). [CrossRef] [PubMed]
30. S. Park, A. B. Karpiouk, S. R. Aglyamov, and S. Y. Emelianov, “Adaptive beamforming for photoacoustic imaging,” Opt. Lett. 33(12), 1291–1293 (2008). [CrossRef] [PubMed]
31. Z. L. Deng, X. Q. Yang, H. Gong, and Q. M. Luo, “Two-dimensional synthetic-aperture focusing technique in photoacoustic microscopy,” J. Appl. Phys. 109(10), 104701 (2011). [CrossRef]
32. C. G. A. Hoelen and F. F. M. de Mul, “A new theoretical approach to photoacoustic signal generation,” J. Acoust. Soc. Am. 106(2), 695–706 (1999). [CrossRef]
33. C. K. Liao, M. L. Li, and P. C. Li, “Optoacoustic imaging with improved synthetic focusing,” Proc. SPIE 5967, 255–262 (2005). [CrossRef]
34. A. F. Frangi, W. J. Niessen, K. L. Vincken, and M. A. Viergever, “Multiscale vessel enhancement filtering, ” in Proceedings of Medical Image Computing & Computer Assisted Intervention (MICCAI), W. Wells, A. Colchester, and S. Delp, eds., 1496 of Lecture Notes in Computer Science, (Springer- Verlag, Berlin, 1998), 130–137.
35. P. F. Hemler, E. S. McCreedy, and M. J. McAuliffe, “Performance evaluation of multiscale vessel enhancement filtering,” Proc. SPIE 5370, 1785–1794 (2004). [CrossRef]
