Improvement of imaging and image correction methods for the soft X-ray projection microscopy

Abstract

BACKGROUND:

The soft X-ray projection microscope has been developed for high resolution imaging of hydrated bio-specimens. Image blurring due to X-ray diffraction can be corrected by an iteration procedure. The correction is not efficient enough for all images, especially for low contrast chromosome images.

OBJECTIVE:

The purpose of this study is to improve X-ray imaging techniques using a finer pinhole and reducing capture time, as well as to improve image correction methods. A method of specimen staining prior to the imaging was tested in order to capture images with high contrasts. The efficiency of the iteration procedure and its combined version with an image enhancement method was also assessed.

METHODS:

In image correction, we used the iteration procedure and its combined version with an image enhancement technique. To capture higher contrast images, we stained chromosome specimens with the Platinum blue (Pt-blue) prior to the imaging.

RESULTS:

The iteration procedure combined with image enhancement corrected the chromosome images with 329 or lower magnification effectively. Using the Pt-blue staining for the chromosome, images with high contrast have been captured and successfully corrected.

CONCLUSIONS:

The image enhancement technique combining contrast enhancement and noise removal together was effective to obtain higher contrast images. As a result, the chromosome images with 329 or lower times magnification were corrected effectively. With Pt-blue staining, chromosome images with contrasts of 2.5 times higher than unstained case could be captured and corrected by the iteration procedure.

Keywords

Soft X-ray projection microscopy imaging Pt-blue image correction iteration procedure contrast enhancement chromosome grayscale value

1 Introduction

The soft X-ray covers a wavelength region called water window where the X-ray absorption by water is significantly smaller than that by organic materials [1, 2]. This gives the possibility to observe the biological specimens in hydrated conditions [3 –5] and becomes the basic reason to build the soft X-ray microscope. In the soft X-ray microscopy, biological specimen does not need drying prior to the imaging as it is needed for the other microscopes using electrons. Also, it gives images with weak radiation damage compared to that by an electron microscope [6 –9].

The simplest scheme for the soft X-ray microscope is the projection type scheme in which a spherical wave from a point source illuminates the object and its magnified image is projected on the detector plane [10]. This type of microscope is also, referred as a shadow or shadow projection microscope in earlier sources [11]. The soft X-ray projection microscope has many advantages such as wide viewing area, easy zooming function, easy extension to CT (Computed Tomography), and simple optical layout particularly, for biological specimens compared to other types of microscopes [10, 12]. But, the projection image is blurred by the diffraction of soft X-rays and contains diffraction fringes around the specimen image and this leads to worsening of spatial resolution [13].

The image of specimens taken by a projection microscope can be corrected (reconstructed) by using the iterative algorithm initiated by R.W.Gerchberg and W.O.Saxton [14] and developed by J.R.Fienup [15] and others [9 , 17]. Hereafter, we use the term iteration procedure instead of the iterative algorithm throughout this article, as we have used in our previous studies. The optical image can be regarded as the Fourier transform of the wave coming on objects [18]. The wave coming from object to detector is registered as a projection image, which is a two dimensional array on a detector and the each element of the array corresponds to only the intensity value on the given pixel. The Fourier domain values, which are the amplitudes on the detector, are taken as the square root of the intensity values. In the iteration procedure, the object domain is regarded as the object wave. The phase information of waves is lost in the image. The initial detector wave is constructed with the amplitude arrays on the detector and guessed phase arrays. This procedure involves the iterative transform back and forth between domains repetitively, with the known constraints for each domain [14, 15]. The detailed consideration of the procedure for our study will be handled in the Section 3. The procedure and its modified versions were successfully used in image reconstruction of various cases [9 , 19–22]. The iteration always converges with the speed depending on constraints and the solution obtained by the procedure is unique for two dimensional case [23]. The procedure was confirmed to be sufficiently effective for soft X-ray projection image reconstruction processes by our earlier studies [13 , 25].

We have improved the imaging conditions and the image correction methods of the soft X-ray projection microscopy in this study. We used a more narrow pinhole with the diameter compared to that used in our previous studies. The narrow pinhole increases the spatial coherency of the wave illuminating the object. Moreover, the higher X-ray intensity was achieved by changing X-ray energy, resulting in the improved contrast compared to the previous study. We stained the chromosome with the Platinum blue (Pt-blue) prior to the imaging to increase the image contrast and in image reconstruction or correction for small objects we used the iteration procedure with image enhancement technique. The image enhancement technique used here consists of contrast enhancement and noise removal techniques together. They were improved to be effective by previous studies separately [24, 26].

2 Objective

Although the iteration procedure works for reconstructing the proper image of objects in most cases, there are cases that iteration procedures are insufficient for image correction and resulted in very poor correction effectiveness [24, 25]. These cases encounter in image correction of thin specimens such as chromosomes, composed basically of light chemical elements. The reason leading to low effectiveness for correction is attributed to the high background noise and the low image contrast. When the contrast of projection images was low, the object was hardly detected by the iteration program and as a result, the images with deformation were created after the iteration procedure. In contrast, the iteration program can work effectively when the contrast of projection images was high. These cases will be called as the images with low and high contrast, respectively in the following texts. Even though applying the image enhancement technique (noise removal and contrast enhancement) to images of latex particle and chromosomes with 329 or lower times magnification gives a suitable image for our iteration procedure, it is not sufficient for chromosome images with magnifications of 504 or higher times.

To overcome this problem, the chromosome specimens were stained with Pt-blue to enhance the contrast of projection images. In addition to preprocessing of the specimen with Pt-blue staining, we also applied an image enhancement technique i.e. noise reduction and contrast enhancement together which were tested previously [24, 25]. The images with Pt-blue staining and image enhancement are acceptable for correction analysis. In this article, we address image correction with iteration procedure of images captured with improved imaging conditions, comparison of results by different procedures from our previous studies [24, 25].

3 Materials and methods

We used a bending magnet beamline BL-11A at Photon Factory of KEK (High Energy Accelerator Research Organization) in Tsukuba, Japan, as a monochromatic soft X-ray source for projection images. The BL-11A beamline provides soft X-rays in the energy ranging from 70 eV to 1900 eV with a grazing-incidence monochromator. In the grazing-incidence monochromator the incident angle is close to π/2, because the reflectivity at short wavelengths increases dramatically when the incident angle increases and monochromatic X-rays are chosen by rotating the gratings based on the reflectivity difference between the X-rays with different energies. The projection experiments were performed using a soft X-ray projection microscope [10, 13] and the selected soft X-ray energies were 700 eV and 1000 eV. Its optical layout and the coloured photo picture are shown in Figs. 1 2, respectively.

Fig. 1

Optical layout of Soft X-ray Projection Microscopy.

Fig. 2

Soft X-ray projection system installed for experiments [27].

In obtaining a point source for spherical wave from the soft X-ray beamline, the projection system in the soft X-ray projection microscope was constructed with a zone-plate and a pinhole made of stainless steel with 1 µm diameter. The new small pinhole was applied in order to have high coherence X-rays generated from a point source in expecting the improvement of spatial resolution and it was located at the focal point of the zone-plate. The zone plate used in the microscope has 0.625 mm diameter with 80 nm of the outermost zone width and 14.142 µm of central block size made of 400 nm thick Tantalum (NTT Advanced Technology Corp., Japan) [10]. The performance of the zone plate was reported in the previous paper [28], which suggests that the real value of the outermost zone width is close to the design value by measuring spatial resolution, because the outermost zone width is responsible for the spatial resolution. The zone-plate and the pinhole were installed in the high vacuum chamber. A back-illuminated X-ray CCD camera with 24.8 µm pixel pitch (Hamamatsu Photonics C4880-30-16W) was used for imaging. For the imaging, a specimen was placed between the pinhole and the CCD camera and magnified images of the specimen were captured on the CCD screen as projection images. The CCD camera and specimen stage were located in the vacuum chamber. The CCD screen was located at the distance of 329 mm from the pinhole. The projection images were captured at magnifications of 47 to 658 times. The experimental parameters are shown in Table 1.

Table 1

Experimental parameters

Parameters	Value
X-ray energy [eV]	700 and 1000
Pinhole diameter [µm]	1
Distance (Pinhole –CCD) [mm]	329
Magnification [times]	47 - 658
Projection time [min]	1 and 3

Latex particles with 10 µm diameter, composed of biopolymer materials of relatively high density (1.03 g/cm³) and chromosomes were used as specimen in this experiment. The latex particle was prepared by drying up the emulsion [24]. Because of their high density, the latex particles give high contrast projection images. Since the latex particles have a purely spherical shapes and definite sizes they were used for comparing their corrected images by the iteration program with our original shape. The chromosome preparation was made from cultured human lymphocyte cell line (IM-9) on a silicon nitride window with 1 mm square in size and 100 nm thickness supported by a center-hollowed silicon substrate with 5 mm square in size and 380 µm in thickness in accordance with the standard protocol for light microscopic observation with some minor modifications [25, 29]. In this study, we also stained the chromosomes by Pt-blue because it is used as common contrast enhancing material for biological specimens [30].

The projection image of specimens captured with soft X-ray projection microscopy is blurred due to diffraction of X-rays. If the phase and amplitude distributions of X-rays on CCD screen were known, the blur correction could be easily made only by using the inverse Fourier Transform (IFT) calculation for inverse propagation of the X-rays to the specimen surface. However, the phase is lost and only the intensity, which is the square of the amplitude module, is registered in the captured picture. To make the blur correction we use the iteration procedure. This procedure performs repeated calculations of the Fourier Transform (FT) and IFT between the specimen surface and CCD screen by using the X-ray amplitude information on the CCD screen extracted from projection images in order to approach unknown phase information of X-ray. Finally, it becomes possible to calculate the X-ray intensity distributions on the specimen surface correctly and results in the corrected image. At the first step of the iteration procedure, the X-ray amplitude distribution data on the CCD screen is combined with phase distribution of spherical wave. Then the real phase distribution was estimated by repetitions of cycled calculations of FT and IFT taking into account of restriction conditions using the following equations:

$F (m T) = \sum_{n = 0}^{N - 1} f (n T_{0}) exp [\frac{i π}{λ R} {(m T - n T_{0})}^{2}],$ (1)

$f (n T_{0}) = \sum_{n = 0}^{N - 1} F (m T) exp [- \frac{i π}{λ R} {(m T - n T_{0})}^{2}],$ (2)

where F (mT)-amplitude distribution on CCD screen, f (nT₀)-amplitude distribution on specimen surface, λ-wavelength, R-distance between specimen and CCD screen, N-sampling number, T₀ = λR/L-sampling number interval on the specimen surface, T-sampling number interval on CCD screen, L- sampling region on the CCD screen, m, n-positive integers. The schematic view of the algorithm for the iteration procedure is shown in Fig. 3. The restriction conditions (constraints) are taken as the same in Ref. [24]. The code implementing the iteration procedure was written in C programming language.

Fig. 3

Schematic view of algorithm for the iteration procedure (RC-Restriction condition for X-ray intensity distribution on the specimen surface).

As mentioned above, the image enhancement method is applied to projected images prior to the iteration procedure. This enhancement method has exactly two steps. In the first, the contrast of projection image improved by contrast enhancement method in which the original grayscale values of the recorded data linearly expands into a new distribution [31]. In other word, the new grayscale distribution for 16-bit images will be from 0 to 65535. In the second step, the background noises of projection image are removed by median filter method. The noise removal is possible without any influence to the image edges by the median filter method which replaces a original grayscale value of a pixel by the median of the grayscale values of the pixels surrounding neighborhood [32].

4 Results and discussion

We have performed three tasks on image correction and compared the results. Firstly, we applied the iteration procedure alone on the images of latex particles and chromosomes with various magnifications. After that, the image enhancement method was applied to projection images to improve the effectiveness of the iteration procedure. Finally, in addition to the image enhancement, the iteration procedure was applied to images of the chromosome stained with Pt-blue.

Effectiveness of the iteration procedure alone. The effectiveness of the correction is evaluated by comparison of shape and size for corrected image with the original object as well as the appearance and disappearance of diffraction fringes. The effectiveness of iteration procedure alone was evaluated on the images of latex particle and chromosomes with magnifications from 47 to 658 times which were captured by using the soft X-rays with 700 eV. The capturing time was 3 minutes. The results of evaluation test confirmed that the iteration procedure is effective for the projection images of latex particles with magnification of 165 times or lower. But, it failed for latex particle images with magnifications of 219 times or higher. The projection and corrected images for the latex particle were shown in Fig. 4 as a successful result. The scaling bars on the corrected and projection images correspond to the real size and magnified size of the specimen, respectively. Because these images are shown in pixels, their sizes look the same. In other word, sampling interval sizes of 0.047 µm on the specimen surface and 0.024 mm on the CCD screen are shown in pixels with size of 0.3 mm for the Fig. 4a,b. The disappearance of diffraction fringes in the projection image is clearly seen on the image in Fig. 4b. As well, the size of object is equal to that for the image after the iteration procedure.

Fig. 4

Images of latex particle (Magnification 165 times). a) Projection image, b) corrected image.

The projection image and its corrected image for chromosome with 219 times magnification were shown in Fig. 5a,b as an unsuccessful result. We examined the grayscale values along red line for projection image and along green line for corrected image. The lines were shown on the images with dashed lines. The result was shown on Fig. 5c. The blue circles on the grayscale distribution shown on Fig. 5c denotes the grayscale maximum values and this implies that the diffraction fringes were not eliminated by the iteration procedure. Actually, the diffraction fringes clearly seen on projection images with 329 times magnification still appeared on the corrected images after applying the iteration procedure. The reason why diffraction fringes did not disappear on the corrected image is supposed due to the poor detectability of the diffraction fringes in the iteration program especially for the low contrast chromosome image.

Fig. 5

Images of a chromosome: (a) projection image, (b) corrected image with the iteration procedure alone (Magnification 219 times). c) Grayscale distributions on a cross-sectional line of chromosome images. The grayscale maximum values were marked by blue circles.

Moreover, the objects on the projection images with 504 or higher magnification for chromosome disappeared partially or fully after the iteration procedure alone (Fig. 6a,b). This result was consistent with the previous results for the evaluation of the noise removal effectiveness prior to the iteration procedure and noise limits on the projection images [24 , 27]. We suppose that this is caused by the low contrast of projection image. In this experiment, the contrast of the images got lowered or almost lost for the chromosome specimen at magnifications higher than 504 times. This contrast lowering or loss was considered to be caused by increase in X-ray scattering and background noises presumably due to too high illumination and too high X-ray intensity on the specimen surface. Thus, we conclude that the application of the iteration procedure alone to chromosomes has some magnification-dependent limitation for the images of chromosome if no strategy of image enhancement is applied.

Fig. 6

Example of unsuccessful correction for a chromosome specimen (The objects on the projection images (a) disappears in the corrected image (b) after the iteration procedure. Magnification 504 times).

Effectiveness of image enhancement method. Since the effectiveness of our image correction depends on image contrasts, we tried to enhance the contrast of images prior to the iteration procedure. In order to increase the contrast of the projection images, an image enhancement method was applied. As a result, the correction effectiveness was greatly improved, and all images of the latex particle and images of the chromosome with 329 times or lower were corrected successfully. We showed the projection image of a chromosome with 329 times magnification as Fig. 7a, the corrected image without the image enhancement (Fig. 7b) and the corrected image with the image enhancement (Fig. 7c).

Fig. 7

Images of a chromosome: (a) projection image, (b) corrected image without image enhancement and (c) corrected image with image enhancement (Magnification 329 times).

Grayscale values along chosen lines (blue and red dashed line) on the images in Fig. 7b,c were given in Fig. 8 correspondingly depicted by blue and red lines. The values corresponding to the diffraction fringe is marked by dashed blue circles. The result of the iteration procedure to the image with the image enhancement is represented by the red line on the figure. Actually, the grayscale values corresponding to fringes are not seen in red lines and this shows the disappearance of the diffraction fringes. However, application results of the iteration procedure show that the iteration procedure is still not effective for images of chromosomes in the high magnification of 504 times or higher. This implies the effectiveness of iteration procedure for the image correction strongly depends on the contrast of the specimen. Quantitative data for contrasts and sizes of the object, and diffraction fringes measured from the correction images for the Fig. 6 Fig. 7 were also compared in Table 2.

Fig. 8

Grayscale distributions on a cross-sectional line of chromosome images. The maximum grayscale values corresponding to the diffraction fringes were marked by blue dashed circles.

Table 2

Contrasts and lengths of the specimen measured from the correction image of chromosome with 329 and 512 times magnifications (Fig. 7 and Fig. 6)

Cases		Contrast, ΔG		Length of specimen [µm]
		Specimen	Diffraction fringe
Chromosome image with magnification of 329 times (Fig. 7)	without image enhancement	10490 ± 55	2512 ± 61	10.2 ± 1.2
	with image enhancement	10253 ± 37	0	9.5 ± 0.1
Chromosome image with magnification of 512 times (Fig. 6)	without image enhancement	0	0	0
	with image enhancement	0	0	0

When the diffraction fringes are not seen on the correction image, the specimen size (length) was measured accurately because of that the edges of the specimen were clearly observed on the image. However, the size was measured with relatively high error when the diffraction fringes remain on the projection image. The specimen sizes for the images corrected with and without image enhancement prior to the iteration procedure were also apparently different. Both of the diffraction fringes and the object are not identifiable for the image with magnification of 504 times due to its too low contrasts. As one of effective ways of contrast enhancement of the specimens, Pt-blue staining method was introduced in the present study.

Effectiveness of Pt-blue staining. The effectiveness of image correction procedure for the images of the specimen stained with Pt-blue was evaluated on the images of chromosome with low contrast which were captured by using the soft X-ray with 1000 eV for 1 minute, because in the beamline used the intensity of monochromatic X-rays was the highest around 1000 eV. The shorter capturing time improves the images with low contrast, because the soft X-ray intensity corresponding to this energy was higher than that in the energy of 700 eV. The projection and corrected images of a chromosome with and without Pt-blue staining were compared in Fig. 9.

Fig. 9

Images of chromosomes with and without Pt-blue. a) Projection image without Pt-blue, b) corrected image without Pt-blue, c) projection image with Pt-blue, d) corrected image with Pt-blue, (Wavelength 1000 eV, Magnification 110 times).

In the case with Pt-blue staining, the image contrast of projection image (Fig. 9c) was apparently higher than that in the case without Pt-blue (Fig. 9a). Applying the correction procedure to images with Pt-blue staining (Fig. 9c) gives a corrected image in which the object morphology became clearer (Fig. 9d) than that in the image (Fig. 9b) corrected from the the projection image without Pt-blue (Fig. 9a). This confirms that the image correction procedure works enough to disappear the diffraction fringes for the images with Pt-blue staining. This results in the successful recognition of chromosome morphology by the iteration program. Hence, it can be concluded that the iteration procedure on the images of the specimen stained with Pt-blue leads almost to the expected result. In testing of image contrast, we calculated the grayscale values of images along the chosen lines drawn by dashed on Fig. 9a,c. The results of the calculation for chromosome was given in Fig. 10.

Fig. 10

Comparison of grayscale distributions along lines on projection images in Fig. 9.

This graph shows the grayscale values versus pixels for the blue and red lines on images, respectively. This result shows that the contrast corresponding to object apparently increased about 2.5 times by the Pt-blue staining. It is represented with ΔG on the graph. Contrasts of the object and diffraction fringes, and object sizes for the correction images were also compared in Table 3. Object sizes and contrasts could be measured only when they are stained with Pt-blue. For the chromosome specimens, their morphology seen in the projection image with the magnification of 658 times was hardly observable by the human eyes (Fig. 11a). However, its morphology became recognizable on the image when the specimen had been stained with Pt-blue (Fig. 11b, white dashed line).

Table 3

Contrasts and lengths of the specimen measured from the correction image of chromosome with 110 times magnification (Fig. 9)

Cases	Contrast, ΔG		Length of specimen [µm]
	Specimen	Diffraction fringe
without Pt-blue	0	0	0
with Pt-blue	320-6700	0	2.1-5.5

Fig. 11

Images of chromosome. a) Projection image without Pt-blue, b) projection image with Pt-blue (Wavelength 1000 eV, Magnification 658 times).

5 Conclusions

The latest research results on the imaging and image correction in the soft X-ray projection microscopy were given in this article. In this study, we aimed to improve the imaging conditions and the image correction methods of soft X-ray projection microscopy. For the improvement of the imaging method, we used a pinhole with diameter of 1µm made of stainless steel. This was finer than that for the previous studies. The narrow pinhole enables us to get a point source producing a significant spherical wave to illuminate the object (specimen) and this increases the spatial coherency of the incident wave. The X-ray intensity was increased by using higher X-ray energy. However, in terms of the X-ray interaction with matter, higher energy increases transmittance to specimens resulting in the lowering image contrast. In the present study our choice was found to be successful because increased intensity also increase in signal to noise ratio, in spite of the decrease of contrast intrinsic in the high energetic X-rays. The spatial coherency of the X-ray wave from the source was acceptable in the limit of pinhole diameters. In this manner, all projection images of the objects were captured with this small (narrow) pinhole. At the present stage, the resolution in this study is not satisfactory compared with other types of X-ray microscopy, because the X-ray intensity from bending magnet beamline was not enough to achieve good image contrast. But the projection type has many advantages over other types of X-ray microscopy as described in Refs. [8 , 13]. Our research can be evaluated as the development of image processing to improve spatial resolution which is inevitable and fundamental process in dealing with projection X-ray images. As well, it should be noted that the projection type soft X-ray microscope is not yet, a properly developed setup. It still includes tasks required to be sophisticated.

The effectiveness of the iteration procedure was evaluated on projection images of chromosomes and latex particles with various magnifications in the beginning of this study. It was effective for the projection images of latex particles with high contrast and magnification of 165 times or lower. However, the image correction was not successful for the images of chromosomes with low contrast at which the object and the diffraction fringe were hardly detected due to its too low contrast and high background noise. To overcome these difficulties, we tested an image enhancement method for the projection images which reduces the image noises and enhances its contrasts, and it was proved that all projection images of the latex particles and the chromosomes with 329 times magnification or lower are correctable. However, the images with magnification of 504 times or higher were not correctable by using the image enhancement method. The application of iteration procedure to those images with enhancement technique results in broken images in which some or whole parts of the specimen image were disappeared due to its too low contrast. The results taken in this research show that using the iteration procedure in retrieving the original image is sufficient for low and medium range magnification image, however it is still not appropriate for high magnification images.

To enhance the image contrast, the chromosome specimens were stained with Pt-blue. The treatment enabled us to capture images with high contrast. As a result, the image contrast was enhanced about 2.5 times (Fig. 10), and the chromosome projection images of low contrast, which were not correctable by the iteration procedure with the image enhancement, were corrected successfully. From the results above, one can conclude that the staining with Pt-blue is the most promising for the X-ray projection microscope because it gives detailed and recognizable images for the correction procedure. Nevertheless, the high magnification images of chromosome remained as unsuccessful targets for the iteration procedure. This indicates that we need to apply another additional technique for the imaging with the projection microscope. This can be phase-contrast imaging [33, 34]. The results obtained in this approach may be beyond from the result taken in other types of X-ray microscope.

In the next step of the study, we are planning to focus on the improvement of the source coherency to increase the imaging resolution for the observation of detailed structures of bio-molecules by using an undulator beamline. This is expected to improve a coherent diffraction imaging technique. In addition, an undulator beamline is also extremely useful in improving image contrast. Much more intense X-rays can be obtained even in the lower energy X-rays, which gives a good signal to noise ratio. Identifying the factors causing decrease in the source coherency is considered a key solution to improve coherency. To accomplish this, studies should be addressed on the influence of size and shape of the source, especially, soft X-ray in the bending magnet beamline and the non-monochromatic spectrum to the source coherency by theoretical calculation with experimental condition and simulation experiment based on imaging systems with LED and laser sources.

Funding

No funds, grants, or other support was received.

Conflict of statement

We have no conflicts of interest to disclose.

Ethical approval

The chromosome used in this research was obtained from open access sources, and it is publicly available.

Informed consent

Informed consent is not applied to this research, because the use of publicly available cell lines has been widely accepted and established in biomedical research.

Author contribution statement

E.J and D.B performed numerical analysis of the pictures using the iteration procedure, designed and written by T.S and E.J. V.J, E.J and D.B wrote the manuscript with input from all authors and contributed to interpretation to the results. A.I, Y.K and T.S performed experiments besides the sample preparation and took the pictures used in the iteration procedure as well contributed to interpret results. All authors reviewed and revised the manuscript.

References

Sayre

, Kirz

, Feder

, Kim

D.M.

and Spiller

, Potential Operating Region for Ultrasoft X-Ray Microscopy of Biological Materials, Science 196(4296) (1977), 1339–1340, doi: 10.1126/science.867033.

Majima

, Soft X-ray imaging of living cells in water: flash contact soft X-ray microscope, TrAC Trends in Analytical Chemistry 23(7) (2004), 520–526, doi: 10.1016/S0165-9936(04)00735-6.

Weigert

, Porat-Shliom

and Amornphimoltham

, Imaging cell biology in live animals: Ready for prime time, Journal of Cell Biology 201(7) (2013), 969–979, doi: 10.1083/jcb.201212130.

Kirz

, Jacobsen

and Howells

, Soft X-ray microscopes and their biological applications, Quarterly Reviews of Biophysics 28(1) (1995), 33–130, doi: 10.1017/S0033583500003139.

Yamamoto

and Shinohara

, Application of X-ray microscopy in analysis of living hydrated cells, The Anatomical Record 269(5) (2002), 217–223, doi: 10.1002/ar.10166.

Lider

V.V.

, X-ray microscopy, Physics-Uspekhi 60(2) (2017), 187–203, doi: 10.3367/UFNe.2016.06.037830.

Williams

, Zhang

, Jacobsen

, Kirz

, Lindaas

, Van’t Hof

and Lamm

S.S.

, Measurements of wet metaphase chromosomes in the scanning transmission X-ray microscope, Journal of Microscopy 170(2) (1993), 155–165, doi: 10.1111/j.1365-2818.1993.tb03335.x.

Shinohara

and Ito

, Radiation damage in soft X-ray microscopy of live mammalian cells, Journal of Microscopy 161(3) (1991), 463–472, doi: 10.1111/j.1365-2818.1991.tb03104.x.

Shen

, Bazarov

and Thibault

, Diffractive imaging of nonperiodic materials with future coherent X-ray sources, Journal of Synchrotron Radiation 11(5) (2004), 432–438, doi: 10.1107/S0909049504016772.

10.

Shinohara

, Ito

, Honda

, Yoshimura

and Yada

, Development of an X-ray projection microscope using synchrotron radiation, AIP Conference Proceedings 507 (2000), 346–349, doi: 10.1063/1.1291169.

11.

Michette

A.G.

, Alan

, Morrison

G.R.

and Buckley

C.J.

, editors, X-ray microscopy III:Proceedings of the Third International Conference: 1990, London, September 3-7, Berlin Heidelberg, Springer-Verlag, c1992. pp. 491.

12.

Mayo

S.C.

, Miller

P.R.

, Wilkins

S.W.

, Davis

T.J.

, Gao

, Gureyev

T.E.

, Paganin

, Parry

D.J.

, Pogany

and Stevenson

A.W.

, Quantitative X-ray projection microscopy: phase-contrast and multispectral imaging, Journal of Microscopy 207(2) (2002), 79–96, doi: 10.1046/j.1365-2818.2002.01046.x.

13.

Shiina

, Suzuki

, Honda

, Ito

, Kinjo

, Yoshimura

, Yada

and Shinohara

, Fresnel diffraction correction by phase-considered iteration procedure in soft X-ray projection microscopy, Journal of Physics: Conference Series (IOP) 186 (2009), 0120591–012059-3, doi: 10.1088/1742-6596/186/1/012059.

14.

Gerchberg

R.W.

and Saxton

W.O.

, A practical algorithm for the determination of phase from image and diffraction plane pictures, Optik 35(2) (1972), 237–246.

15.

Fienup

J.R.

, Phase retrieval algorithms: a comparison, Appl Opt 21(15) (1982), 2758–2769, doi: 10.1364/AO.21.002758.

16.

Miao

, Sayre

and Chapman

H.N.

, Phase retrieval from the magnitude of the Fourier transforms of nonperiodic objects, J Opt Soc Am A 15(6) (1998), 1662–1669, doi: 10.1364/JOSAA.15.001662.

17.

Marchesini

, He

, Chapman

H.N.

, Hau-Riege

S.P.

, Noy

, Howells

M.R.

, Weierstall

and Spence

J.C.H.

, X-ray image reconstruction from a diffraction pattern alone, Phys Rev B 68(14) (2003), 1401011–140101-4, doi: 10.1103/PhysRevB.68.140101.

18.

Born

and Wolf

, Principles of Optics: 60th Anniversary Edition, 7th Edition, London, Cambridge University Press, 2019.

19.

Fienup

J.R.

, Reconstruction of an object from the modulus of its Fourier transform, Opt Lett 3(1) (1978), 27–29, doi: 10.1364/OL.3.000027.

20.

Latychevskaia

, Reconstruction of missing information in diffraction patterns and holograms by iterative phase retrieval, Optics Communications 452 (2019), 56–67, doi: 10.1016/j.optcom.2019.07.021.

21.

Hemonnot

C.Y.J.

and Koster

, Imaging of Biological Materials and Cells by X-ray Scattering and Diffraction, ACS Nano 11(9) (2017), 8542–8559, doi: 10.1021/acsnano.7b03447.

22.

Shapiro

, Thibault

, Beetz

, Elser

, Howells

, Jacobsen

, Kirz

, Lima

, Miao

, Neiman

A.M.

and Sayre

, Biological imaging by soft X-ray diffraction microscopy, Proceedings of the National Academy of Sciences 102(43) (2005), 15343–15346, doi: 10.1073/pnas.0503305102.

23.

Seldin

J.H.

and Fienup

J.R.

, Numerical investigation of the uniqueness of phase retrieval, J Opt Soc Am A 7(3) (1990), 412–427, doi: 10.1364/JOSAA.7.000412.

24.

Jamsranjav

, Kuge

, Ito

, Kinjo

and Shiina

, Effectiveness of noise removal for image correction in soft X-ray projection microscopy, Radioisotopes 66(4) (2017), 137–148, doi: 10.3769/radioisotopes.66.137.

25.

Jamsranjav

, Kuge

, Ito

, Kinjo

and Shiina

, Evaluation of noise limits to improve image processing in soft X-ray projection microscopy, Journal of X-ray Science and Technology 25(4) (2017), 653–665, doi: 10.3233/XST-16201.

26.

Jamsranjav

, Shiina

, Kuge

, Kinjo

, Nakamura

, Shinohara

and Ito

, Effect of contrast enhancement prior to iteration procedure on image correction for soft X-ray projection microscopy, AIP Conference Proceedings 1696 (2016), 0200371–020037-5, doi: 10.1063/1.4937531.

27.

Jamsranjav

, Image correction for soft X-ray projection microscopy (Doctoral dissertation), Chiba University, Japan, 2017.

28.

Ito

, Shinohara

, Honda

, Saito

, Kuriyama

, Yoshimura

, Maeda

, Hori

, Endoh

, Komatsubara

, Achihara

and Kinjo

, Improvement of spatial resolution in soft X-ray projection microscope, Photon Factory Activity Report 19 (2001), 208.

29.

Liehr

and Claussen

, FISH on Chromosome Preparations of Peripheral Blood. In: B. Rautenstrauss and T. Liehr editors, FISH technology, Springer lab manual, Heidelberg Berlin, Springer Verlag, 2002, pp. 73–81.

30.

Inaga

, Katsumoto

, Tanaka

, Kameie

, Nakane

and Tomonori

, Platinum blue as an alternative to uranyl scetate for staining in transmission electron microscopy, Archives of Histology and Cytology 70(5) (2007), 43–49, doi: 10.1679/aohc.70.43.

31.

Al-amri

S.S.

, Kalyankar

N.V.

and Khamitkar

S.D.

, Linear and non-linear contrast enhancement image, IJCSNS International Journal of Computer Science and Network Security 10(2) (2010), 139–143.

32.

Baxes

A.G.

, Digital image processing: Principles & applications, New York, Wiley & Sons Inc., 1994.

33.

Tao

, He

, Hao

, Kuang

and Liu

, Principles of Different X-ray Phase-Contrast Imaging: A Review, Applied Sciences 11(7) (2021), 2971, doi: 10.3390/app11072971.

34.

Nakasako

, Kobayashi

, Takayama

, Asakura

, Oide

, Okajima

, Oroguchi

and Yamamoto

, Methods and application of coherent X-ray diffraction imaging of noncrystalline particles, Biophysical Reviews 12(2) (2020), 541–567, doi: 10.1007/s12551-020-00690-9.