Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Jul 12.
Published in final edited form as: Appl Opt. 2020 Aug 1;59(22):G41–G46. doi: 10.1364/AO.392004

Speckle-free, near-infrared portable confocal microscope

Cheng Gong 1, Delaney B Stratton 2, Clara N Curiel-Lewandrowski 2,3, Dongkyun Kang 1,3,4,*
PMCID: PMC8273882  NIHMSID: NIHMS1719018  PMID: 32749315

Abstract

We have developed a portable confocal microscope (PCM) that uses an inexpensive near-infrared (NIR) LED as the light source. Use of the spatially incoherent light source significantly reduced the speckle contrast. The PCM device was manufactured at the material cost of approximately $5,000 and weighed only 1 kg. Lateral and axial resolutions were measured as 1.6 μm and 6.0 μm, respectively. Preliminary in vivo skin imaging experiment results showed that the PCM device could visualize characteristic cellular features of human skin extending from the stratum corneum to the superficial dermis. Dynamic imaging of blood flow in vivo was also demonstrated. The capability to visualize cellular features up to the superficial dermis are expected to facilitate evaluation and clinical adoption of this low-cost diagnostic imaging tool.

1. Introduction

Reflectance confocal microscopy (RCM) is a non-invasive optical imaging method that can examine cellular details of the skin with a quasi-histologic resolution [1–5]. Through numerous clinical studies, RCM has been shown to diagnose the vast majority of skin cancers with high sensitivity and specificity [6–11], which significantly surpass those of the clinical assessment. Recent studies show early evidence of how RCM can improve the skin cancer diagnosis and treatment: i) unnecessary biopsy of benign lesions can be reduced by 50-68% [12, 13]; ii) treatment can be initiated during the initial visit without having to wait for the histologic diagnosis [14]; and iii) non-invasive treatment can be monitored at cellular resolution [15, 16].

In 2016, current procedural terminology (CPT) reimbursement codes have been granted for RCM imaging of the skin [17]. Wide clinical uptake of RCM, however, has been hampered by the following factors: i) the device cost is high, ~$100,000; ii) training is required for reading confocal images; iii) the standard commercial RCM device (Vivascope 1500) is bulky [18]; and iv) the more portable commercial RCM device (Vivascope 3000) has a small field of view (FOV) without dermoscopic guidance resulting in sub-optimal diagnostic accuracy [19]. In response to the barrier associated with training requirement to interpret the RCM images, promising results have been recently reported from development of automated RCM image analysis algorithms [20–22]. It is anticipated that either remote live interpretation or AI-based analysis will aid less experienced healthcare providers interpret confocal images. Major improvement of the RCM hardware to make the device portable and affordable, however, has been challenging.

Previously, we have reported on the development of a low-cost, smartphone-based confocal microscope [23]. In the smartphone confocal microscope, scan-free confocal optics was used, where each wavelength of the source is encoded with a transverse coordinate of the tissue. Since inexpensive optoelectrical components are incorporated in the smartphone confocal microscope, the overall material cost of the device was approximately $4,000. While the smartphone confocal microscope successfully visualized known RCM cellular features of human skin in vivo, there were several technological constraints: the imaging depth was limited due to the use of visible illumination light (central wavelength = 595 nm) and the imaging speed was relatively slow, 4.3 frames/sec (fps), which made the acquired confocal images prone to have motion artifacts.

More recently, we developed a near-infrared (NIR) portable confocal microscope (PCM) to address the aforementioned challenges found in the smartphone confocal microscope [24]. In the previous PCM device, a super-luminescent LED (sLED; central wavelength = 840nm; bandwidth = 50nm) was used with the goal of increasing the imaging depth and speed. While the sLED-based PCM device achieved a very high imaging speed of 203 fps, image quality was noticeably degraded due to the speckle noise. The speckle noise was generated by the use of a spatially-coherent light source (sLED) and detection of narrow spectral bandwidth (0.66 nm) by each pixel. The degraded image quality made it challenging to examine the RCM cellular features.

In this paper, we present a new PCM device that significantly reduces the speckle noise, making it possible to analyze the cellular structures from epidermis to superficial dermis. The new PCM device uses a spatially-incoherent light source, an NIR LED. Specific design of the LED-based PCM device is described. Process and results of simulating theoretical resolution are presented. Preliminary confocal images of human skin in vivo are presented and compared with the images taken from the previous sLED-based PCM device.

2. Methods

2.1. LED-based Portable confocal microscope

An LED (part number = SFH 4780S, Osram.; emission area = 0.75 × 0.75 mm2; viewing angle = 20°; maximum optical power = 660 mW) was used as the light source. The source central wavelength was 820 nm, and 40%-of-maximum bandwidth 40 nm. An aspherical collimator (f = 8 mm) and a cylindrical lens (f = 7.7 mm) were used to focus the light from the LED on an illumination slit (width = 25 μm; length = 3 mm). An achromatic doublet (f = 30 mm) was used to collimate the light from the illumination slit. The collimated beam was diffracted by a transmission grating (groove density = 1765 lpmm), which generated the full field angle of 5.9 ° for the spectral bandwidth of 40 nm. The diffracted light passed through a D-shaped aperture (inset, Fig. 1; width = 3.5 mm; height = 7.9 mm). A water-immersion objective lens (40x; NA = 0.8) was used to focus the illumination light on the tissue at the nominal incidence angle of 18.7° and effective NA of 0.35. The 5.9° field angle corresponded to the field size of 514 μm.

Figure 1.

Figure 1

Open in a new tab

Schematic of the NIR PCM device. black arrow – illumination; red arrow – detection.

Light scattered back from the cellular structure of the tissue was collected by the same objective lens. Another D-shaped aperture was used for the detection beam path, which limited the effective detection NA to 0.35. The divided pupil approach, where the illumination and detection beam paths use separate regions of the pupil, was used to provide higher image contrast and reduce the contribution of the specular reflection from optical elements on the confocal image [25] . After the D-shaped aperture, the detection beam was diffracted by another transmission grating (1765 lpmm) and focused by an achromatic doublet (f = 30 mm) onto a detection slit (width = 25 μm; length = 3 mm). Light filtered by the detection slit was collimated by an achromatic doublet (f = 30 mm), diffracted by a transmission grating (1765 lpmm), and focused by a multi-element camera lens (f = 50 mm) onto a CMOS sensor (acA1300-200um, Basler; 1,280 × 1,024 pixels; pixel size = 4.8 μm). A fold mirror was added in the detection beam path to position the camera lens and CMOS sensor away from the tissue, which facilitated placement of the device on skin lesions. Most of the optical elements were assembled passively using custom, 3D-printed mechanical holders (printed by Form 3, Formlabs and uPrint SE Plus, Stratasys). The detection slit was mounted on a miniature, 3-axis translation stage to make the detection slit conjugate to the illumination slit and achieve proper resolution.

2.2. Resolution simulation

We have simulated the resolution of PCM by calculating the product of the illumination and detection point-spread functions (PSFs). Resolution simulation process is illustrated in Fig. 2. First, the 3D diffraction-limited PSF for the effective NA of 0.35 (water immersion) was calculated using the PSF Generator plug in in ImageJ [26]. The diffraction-limited 3D PSF was rotated by 18.7° (step 1) to represent the illumination beam that was introduced through the left half of the objective lens pupil (inset, Fig. 1) and incident on the tissue at the nominal angle of 18.7°. The rotated PSF was convoluted with the image of the illumination slit on the tissue (step 2). With the magnification of 6 between the illumination slit and tissue, the size of the illumination slit image on the tissue was 4.2 (width) μm × 500 μm (length). The convoluted PSF showed the illumination PSF for a single wavelength when a 25-μm-wide illumination slit is used.

Figure 2.

Figure 2

Open in a new tab

Process for simulating the PCM resolution and resulting confocal PSF and its cross sections.

Next, we analyzed the effective spectrum each pixel was detecting. As was the case in the previous PCM device, each pixel on the CMOS sensor was conjugate to a single point on the tissue. The main role of the detection optics was to limit the spectrum that each pixel detected, which limited the width of the effective illumination beam. Given the groove density of the grating, 1765 lpmm, focal length of the focusing lens, 30 mm, and detection slit width, 25 μm, the bandwidth of the spectrum that each pixel detects was 0.33 nm. The 0.33 nm bandwidth corresponded to a 4.2-μm lateral spread of the illumination beam. The effective illumination PSF was finally calculated by the convolution between the monochromatic illumination PSF for the 25 μm slit with the line width of 4.2 μm (step 3).

Detection PSF was calculated by rotating the diffraction-limited PSF for NA of 0.35 by −18.7° (step 1) to represent the detection beam that was collected through the right half of the objective lens pupil (inset, Fig. 1). The rotated PSF was convoluted with the image of each CMOS sensor pixel on the tissue (step 2). The pixel size of the CMOS sensor was 4.8 μm, and the magnification between the CMOS sensor and tissue was 10, which made the width of the pixel size on the tissue as 0.48 μm. Since the image of the individual pixel had a smaller width than the diffraction-limited PSF width, the final detection PSF was not widened noticeably. The confocal PSF was calculated as the product of the illumination and detection PSFs. The cross sections of the confocal PSF were analyzed along the lateral (x) and axial (z) coordinates of the objective lens and minor (u) and major (v) axes of the PSF. FWHM values were calculated as 1.2 μm (x), 3.6 μm (z), 1.1 μm (u), and 7.2 μm (v).

2.3. Performance test

Lateral resolution was tested by imaging the USAF resolution target. The smallest pattern distinguished was analyzed. FWHM of the line-spread function was calculated along the horizontal and vertical directions of the confocal image. The axial resolution was calculated by imaging a mirror while translating the PCM device axially with a motorized stage. FWHM of the axial response curve was calculated.

Tissue imaging performance was tested by imaging human skin in vivo. During tissue imaging, the exposure time of the CMOS sensor was set as 0.05 second, and the corresponding imaging speed was 20 fps. The confocal images were transferred to a laptop (Surface Book Pro, Microsoft) via the USB 3.0 cable. Data were saved in real time using a custom LabVIEW code. The raw image data were compensated for the intensity non-uniformity using a custom Matlab code: A reference image was generated by i) acquiring a tissue image at the imaging depth larger than a typical confocal imaging depth limit, ~ 250 μm, so that the image does not exhibit any microscopic details but shows blurred overall reflectivity of the tissue, and ii) blurring the tissue image further using a Gaussian blurring filter with radius of 20 pixels. The reference image was normalized and used to divide each raw confocal image.

When imaging forearm and finger, the PCM device was mounted on a motorized stage to image the tissue for the imaging depth range of 0 to 300 μm. For the confocal images of forearm, we compared the images obtained by the current LED-based PCM device with the images of the same forearm acquired by the previous sLED-based device. Speckle noise contrast was measured for the image obtained from dermis. Four regions of interest (ROIs) with the area of approximately 4,000 pixels per each ROI that exhibit grossly uniform structure were chosen. Average and standard deviation of the intensity values within each ROI were calculated. The speckle contrast was determined as the ratio between the standard deviation and average. The PCM device was then unmounted from the motorized stage and held by a single hand to image lower lip in vivo. The pressure on the tissue was gently varied to change the imaging depth.

3. Results

The NIR PCM device (Fig. 3a) was manufactured at the material cost of approximately $5,000. The PCM device had a dimension of 22 cm × 17.5 cm × 10 cm, and weighed about 1 kg. The illumination power on the tissue was 0.158 mW. The USAF resolution target image revealed that the pattern in the group 9, element 1 (linewidth = 0.98 μm) is distinguishable along the horizontal direction. The pattern in the group 9, element 3 (linewidth = 0.78 μm) was distinguished along the vertical direction, or the slit length direction. The FWHM of the line spread function (LSF) was measured as 1.68 ± 0.16 μm and 1.55 ± 0.13 μm along the horizontal and vertical directions, respectively. The average resolution is 1.62 μm. The measured FWHM of the axial response curve was 6.01 ± 0.11 μm.

Figure 3.

Figure 3

Open in a new tab

Photo of the PCM device (A) and confocal image of the USAF resolution target (B).

Confocal images of a human forearm obtained at three imaging depths are shown in Figs. 4a–c. At the superficial region with the imaging depth of 26 μm (Fig. 4a), keratinocytes in epidermis including the stratum granulosum and spinosum are well visualized with dark cytoplasm and bright cell borders (arrows in Fig. 4a). At a larger imaging depth of 72 μm (Fig. 4b), dermal papillae (asterisk in Fig. 4b) are delineated by circular arrangement of melanocytes or melanin-containing basal cells (arrows in Fig. 4b). At the imaging depth of 232 μm (Fig. 4c), collagen fibers (arrowheads in Fig. 4c) in dermis are visualized. Speckle contrast in the dermis image (Fig. 4c) was measured as 0.08 ± 0.01, which was significantly lower than those values measured for the previous sLED-based device, 0.24-0.26.

Figure 4.

Figure 4

Open in a new tab

Confocal images of human forearm in vivo obtained with the current LED-based PCM device (A-C) and previous sLED-based device (D-F). Different tissue layers are visualized: epidermis (A, D), dermal-epidermal junction (B, E), and superficial dermis (C, F). arrows in A, D – keratinocytes; asterisks in B, E – dermal papillae; arrows in B – melanocytes or melanin-containing basal cells; arrowheads in C – collagen fibers.

Images taken with the previous sLED-based device are shown in Figs. 4d–f. While some of the cellular features shown in Figs. 4a–c are also observed in Figs. 4d–f such as keratinocytes (arrows in Fig. 4d) and dermal papillae (Fig. 4e), other morphologic features were difficult to appreciate. This comparison shows that the reduction of the speckle noise and improvement of the axial resolution, from 11 μm to 6 μm, in the current LED-based device provides significantly better visualization of cellular features.

Images of the finger are shown in Fig. 5. In the superficial region (Fig. 5a), numerous keratinized cells (arrowheads in Fig. 5a) are shown. At the base of the epidermis (Fig. 5b), regular honeycomb pattern of cells (arrows in Fig. 5b) are observed, which might correspond to the single layer of basal generative cells. Under the basement membrane (Fig. 5c), dermal papillae are visualized as regularly-spaced dark openings (asterisk in Fig. 5c).

Figure 5.

Figure 5

Open in a new tab

Confocal image of human finger in vivo obtained with the LED-based PCM device. Different tissue layers are visualized: epidermis (A), basement membrane (B), and superficial dermis (C). arrowheads – hyper-keratinization; arrows – basal cells; and asterisk – dermal papillae.

In the confocal images of the lower lip (Fig. 6), various structures are revealed. When the pressure on the tissue was light to make the imaging depth superficial, epithelial cell nuclei are visualized as bright dots (Fig. 6a). When the pressure on the tissue was increased to image the basal layer (Fig. 6b), blood cells inside capillaries are visualized (arrows in Fig. 6b, Visualization 1). At a deeper imaging depth inside the lamina propria (Fig. 6b), thin collagen fibers are clearly visualized (arrowheads in Fig. 6c).

Figure 6.

Figure 6

Open in a new tab

Confocal images of human lower lip in vivo obtained with the LED-based PCM device. Different tissue layers are visualized: squamous epithelium (A), basal layer (B), and lamina propria (C). arrows in A– stratified squamous epithelial cell nuclei; arrows in B – red bloods cells inside a capillary (dynamic flow is shown in Visualization 1); and arrowheads – collagen fibers.

4. Conclusions

In this paper, we have presented a NIR PCM and demonstrated cellular imaging of human skin in vivo. Compared to the previous sLED-based device, the current LED-based device significantly reduced the speckle noise, which facilitated the identification of cellular features in confocal images. Additionally, cellular features visualized in the present PCM device were similar to those presented in previous RCM literatures. This might suggest that the LED-based PCM device merit further evaluation for visualizing previously-validated RCM features associated with various skin diseases.

The material cost of our PCM device, ~$5,000, is higher than other smartphone-based portable microscopy devices. However, it is noted that the PCM device provides a unique capability of imaging cellular details of the skin in vivo, which might justify the use of the PCM device in certain resource-limited settings. In addition, we expect that the device cost will be significantly decreased in future commercialization by developing low-cost objective lenses and batch-producing optical elements. Compared to our previous smartphone-based PCM device, our current PCM device provided better image contrast and enhanced imaging depth. The imaging performance improvement was achieved by the use of the NIR-based LED and the monochromatic CMOS sensor without the Bayer color filter. Use of a smartphone as part of the PCM device, however, provides some advantages: reducing the device cost and obviating the need for a laptop for data acquisition. In future development, we will explore a NIR-based, smartphone confocal microscope by removing filters from the smartphone camera module.

Acknowledgment

This research was supported in part by the National Institutes of Health (R21TW010221).

Funding

National Institutes of Health/Fogarty International Center (R21TW010221).

Footnotes

Disclosures

CG and DK are the inventors of the provisional patent application related to the portable confocal microscopy technology presented in this paper. The University of Arizona has a technology-licensing agreement with a start-up company on the presented technology. CG and DK have rights to receive royalties as a result of this licensing agreement.

References

  • 1.Milind Rajadhyaksha MG, Esterowitz Dina, Webb Robert H., Rox Anderson R, “In vivo confocal scanning laser microscopy of human skin: Melanin provides strong contrast,” J. Invest. Dermatol Vol.104, 946–952 (1995). [DOI] [PubMed] [Google Scholar]
  • 2.Milind Rajadhyaksha SG, Zavislan James M., Rox Anderson R, Webb Robert H., “In vivo confocal scanning laser microscopy of human skin II : Advances in Instrumentation and comparison with histology,” J. Invest. Dermatol Vol.113, 293–303 (1999). [DOI] [PubMed] [Google Scholar]
  • 3.Rajadhyaksha M, Anderson RR, and Webb RH, “Video-rate confocal scanning laser microscope for imaging human tissues in vivo,” Applied optics 38, 2105–2115 (1999). [DOI] [PubMed] [Google Scholar]
  • 4.Agozzino M, Moscarella E, Babino G, Caccavale S, Piccolo V, and Argenziano G, “The use of in vivo reflectance confocal microscopy for the diagnosis of melanoma,” Expert review of anticancer therapy (2019). [DOI] [PubMed] [Google Scholar]
  • 5.Dinnes J, Deeks JJ, Chuchu N, Saleh D, Bayliss SE, Takwoingi Y, Davenport C, Patel L, Matin RN, and O’sullivan C, “Reflectance confocal microscopy for diagnosing keratinocyte skin cancers in adults,” Cochrane Database of Systematic Reviews (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Pellacani G, Guitera P, Longo C, Avramidis M, Seidenari S, and Menzies S, “The impact of in vivo reflectance confocal microscopy for the diagnostic accuracy of melanoma and equivocal melanocytic lesions,” Journal of investigative dermatology 127, 2759–2765 (2007). [DOI] [PubMed] [Google Scholar]
  • 7.Gerger A, Hofmann-Wellenhof R, Langsenlehner U, Richtig E, Koller S, Weger W, Ahlgrimm-Siess V, Horn M, Samonigg H, and Smolle J, “In vivo confocal laser scanning microscopy of melanocytic skin tumours: diagnostic applicability using unselected tumour images,” British Journal of Dermatology 158, 329–333 (2008). [DOI] [PubMed] [Google Scholar]
  • 8.Guitera P, Pellacani G, Longo C, Seidenari S, Avramidis M, and Menzies SW, “In vivo reflectance confocal microscopy enhances secondary evaluation of melanocytic lesions,” Journal of investigative dermatology 129, 131–138 (2009). [DOI] [PubMed] [Google Scholar]
  • 9.Guitera P, Pellacani G, Crotty KA, Scolyer RA, Li L-XL, Bassoli S, Vinceti M, Rabinovitz H, Longo C, and Menzies SW, “The impact of in vivo reflectance confocal microscopy on the diagnostic accuracy of lentigo maligna and equivocal pigmented and nonpigmented macules of the face,” Journal of Investigative Dermatology 130, 2080–2091 (2010). [DOI] [PubMed] [Google Scholar]
  • 10.Guitera P, Menzies SW, Longo C, Cesinaro AM, Scolyer RA, and Pellacani G, “In vivo confocal microscopy for diagnosis of melanoma and basal cell carcinoma using a two-step method: analysis of 710 consecutive clinically equivocal cases,” Journal of investigative dermatology 132, 2386–2394 (2012). [DOI] [PubMed] [Google Scholar]
  • 11.Nori S, Rius-Díaz F, Cuevas J, Goldgeier M, Jaen P, Torres A, and González S, “Sensitivity and specificity of reflectance-mode confocal microscopy for in vivo diagnosis of basal cell carcinoma: a multicenter study,” Journal of the American Academy of Dermatology 51, 923–930 (2004). [DOI] [PubMed] [Google Scholar]
  • 12.Pellacani G, Pepe P, Casari A, and Longo C, “Reflectance confocal microscopy as a second-level examination in skin oncology improves diagnostic accuracy and saves unnecessary excisions: a longitudinal prospective study,” The British journal of dermatology 171, 1044–1051 (2014). [DOI] [PubMed] [Google Scholar]
  • 13.Alarcon I, Carrera C, Palou J, Alos L, Malvehy J, and Puig S, “Impact of in vivo reflectance confocal microscopy on the number needed to treat melanoma in doubtful lesions,” The British journal of dermatology 170, 802–808 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Kadouch D, Elshot Y, Zupan-Kajcovski B, van Haersma de With A, van der Wal A, Leeflang M, Jóźwiak K, Wolkerstorfer A, Bekkenk M, and Spuls P, “One-stop-shop with confocal microscopy imaging vs. standard care for surgical treatment of basal cell carcinoma: an open-label, noninferiority, randomized controlled multicentre trial,” British Journal of Dermatology 177, 735–741 (2017). [DOI] [PubMed] [Google Scholar]
  • 15.Ulrich M, Lange-Asschenfeldt S, and Gonzalez S, “The use of reflectance confocal microscopy for monitoring response to therapy of skin malignancies,” Dermatol Pract Concept 2, 202a210 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Chen CS, Sierra H, Cordova M, and Rajadhyaksha M, “Confocal microscopy-guided laser ablation for superficial and early nodular Basal cell carcinoma: a promising surgical alternative for superficial skin cancers,” JAMA Dermatol 150, 994–998 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Rajadhyaksha M, Marghoob A, Rossi A, Halpern AC, and Nehal KS, “Reflectance confocal microscopy of skin in vivo: From bench to bedside,” Lasers in surgery and medicine 49, 7–19 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Rajadhyaksha M, “Confocal microscopy of skin cancers: translational advances toward clinical utility,” in Engineering in Medicine and Biology Society, 2009. EMBC 2009. Annual International Conference of the IEEE, (IEEE, 2009), 3231–3233. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Jain M, Pulijal SV, Rajadhyaksha M, Halpern AC, and Gonzalez S, “Evaluation of Bedside Diagnostic Accuracy, Learning Curve, and Challenges for a Novice Reflectance Confocal Microscopy Reader for Skin Cancer Detection In Vivo,” JAMA Dermatol 154, 962–965 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Kurugol S, Kose K, Park B, Dy JG, Brooks DH, and Rajadhyaksha M, “Automated delineation of dermal–epidermal junction in reflectance confocal microscopy image stacks of human skin,” Journal of Investigative Dermatology 135, 710–717 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Koller S, Wiltgen M, Ahlgrimm-Siess V, Weger W, Hofmann-Wellenhof R, Richtig E, Smolle J, and Gerger A, “In vivo reflectance confocal microscopy: automated diagnostic image analysis of melanocytic skin tumours,” Journal of the European Academy of Dermatology and Venereology 25, 554–558 (2011). [DOI] [PubMed] [Google Scholar]
  • 22.Bozkurt A, Kose K, Alessi-Fox C, Gill M, Dy J, Brooks D, and Rajadhyaksha M, “A Multiresolution Convolutional Neural Network with Partial Label Training for Annotating Reflectance Confocal Microscopy Images of Skin,” in International Conference on Medical Image Computing and Computer-Assisted Intervention, (Springer, 2018), 292–299. [Google Scholar]
  • 23.Freeman EE, Semeere A, Osman H, Peterson G, Rajadhyaksha M, González S, Martin JN, Anderson RR, Tearney GJ, and Kang D, “Smartphone confocal microscopy for imaging cellular structures in human skin in vivo,” Biomed. Opt. Express 9, 1906–1915 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Gong C, Kulkarni N, Zhu W, Nguyen CD, Curiel-Lewandrowski C, and Kang D, “Low-cost, high-speed near infrared reflectance confocal microscope,” Biomed. Opt. Express 10, 3497–3505 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Dwyer PJ, DiMarzio CA, and Rajadhyaksha M, “Confocal theta line-scanning microscope for imaging human tissues,” Applied optics 46, 1843–1851 (2007). [DOI] [PubMed] [Google Scholar]
  • 26.Kirshner H, Aguet F, Sage D, and Unser M, “3-D PSF fitting for fluorescence microscopy: implementation and localization application,” Journal of microscopy 249, 13–25 (2013). [DOI] [PubMed] [Google Scholar]

RESOURCES