Method for Imaging Objects Contained in a Droplet

Abstract

Examples include imaging one or more objects contained inside a droplet. A method includes generating at least one hologram of the one or more objects contained in the droplet by using in-line lens-free imaging. The at least one hologram includes at least one artifact that is caused by the droplet and that affects the at least one characteristic of the one or more objects contained in the droplet. The method includes at least partially removing the at least one artifact or the cause of the at least one artifact. The method further includes generating an image, after or during removing the at least one artifact or the cause of the at least one artifact. The image includes the one or more objects. The method also comprises recognizing the at least one characteristic of the one or more objects based on the image.

Claims

1. A method comprising: generating a hologram of an object contained in a droplet, wherein the hologram is generated using in-line lens-free imaging and includes an artifact that is caused by the droplet and that affects a characteristic of the object contained in the droplet; at least partially removing the artifact or a cause of the artifact from the hologram; thereafter using the hologram to generate an image comprising the object; and determining the characteristic of the object based on the image.

2. The method according to claim 1, wherein the artifact comprises an interference in the hologram of a holographic signal of the droplet.

3. The method according to claim 1, wherein the artifact comprises an interference in the hologram of a holographic signal of the object.

4. The method according to claim 1, wherein the artifact comprises noise in a holographic signal of the droplet.

5. The method according to claim 1, wherein the artifact comprises noise in a holographic signal of the object.

6. The method according to claim 1, wherein at least partially removing the artifact or the cause of the artifact comprises regenerating the hologram of the object, wherein a second illumination wavefront is used for regenerating the hologram that is different from a first illumination wavefront that was used for generating the hologram.

7. The method according to claim 6, wherein the second illumination wavefront comprises a single wavelength.

8. The method according to claim 6, wherein the second illumination wavefront comprises multiple wavelengths.

9. The method according to claim 1, wherein at least partially removing the artifact or the cause of the artifact comprises processing the hologram with a numerical optimization algorithm.

10. The method of claim 9, wherein the numerical optimization algorithm comprises a fast iterative shrinkage-thresholding algorithm.

11. The method of claim 9, wherein the numerical optimization algorithm comprises an Alternating Direction Method of Multipliers algorithm.

12. The method according to claim 1, wherein at least partially removing the artifact or the cause of the artifact comprises processing the hologram with an image processing method.

13. The method according to claim 1, wherein generating the image comprises generating the image by multi-depth reconstruction of the hologram.

14. The method according to claim 1, further comprising: determining a strength of the artifact; and selecting a process for at least partially removing the artifact or the cause of the artifact based on the strength of the artifact.

15. The method according to claim 14, wherein determining the strength of the artifact comprises comparing first parameters of the object in the hologram and second parameters of the object in a reference hologram.

16. The method according to claim 1, wherein the object comprises one or more objects and the characteristic of the one or more objects comprises: a number of the one or more objects contained in the droplet; a density of the one or more objects in the droplet; a respective size and/or shape of each object of the one or more objects in the droplet; a respective position of each object of the one or more objects in the droplet; a motion pattern of the one or more objects in the droplet; or a reflective index of the one or more objects in the droplet.

17. The method of claim 1, wherein at least partially removing the artifact or the cause of the artifact comprises using a computational model to at least partially remove the artifact or the cause of the artifact.

18. The method of claim 17, wherein the computational model comprises a deep learning neural network.

19. A computer program comprising instructions that, when executed by a processor, causes the processor to perform the method according to claim 1.

20. An apparatus comprising: a holographic imager configured to generate a hologram of an object contained in a droplet using in-line lens-free imaging, wherein the hologram includes an artifact that is caused by the droplet and that affects a characteristic of the object in the droplet; and a processor configured to at least partially remove the artifact or a cause of the artifact from the hologram, wherein the processor is further configured to generate an image comprising the object after partially removing the artifact or the cause of the artifact and to determine the characteristic of the object based on the image.

Description

BRIEF DESCRIPTION OF THE FIGURES

[0042] The above, as well as additional, features will be better understood through the following illustrative and non-limiting detailed description of example embodiments, with reference to the appended drawings.

[0043] FIG. 1 schematically illustrates holographic imaging of an object contained in a droplet, according to an example.

[0044] FIG. 2 shows a method for imaging and recognizing at least one characteristic of an object contained in a droplet, according to an example.

[0045] FIG. 3 schematically illustrates different mechanisms of artifacts caused by the presence of a droplet, according to an example.

[0046] FIG. 4 shows a pipeline for removing one or more artifacts or a cause of the one or more artifacts from a hologram, according to an example.

[0047] FIG. 5 compares various methods of removing the one or more artifacts or the cause of the one or more artifacts from the hologram with respect to a strength of improvement and effort, according to an example.

[0048] FIG. 6 shows a method using a FISTA algorithm for removing stronger artifacts from a hologram, according to an example.

[0049] FIG. 7 shows a method using conventional image and signal processing for removing weaker artifacts from a hologram, according to an example.

[0050] FIG. 8a illustrates a method using wavefront engineering to remove causes of artifacts in a hologram, according to an example.

[0051] FIG. 8b illustrates a method using wavefront engineering to remove causes of artifacts in a hologram, according to an example.

[0052] FIG. 8c illustrates a method using wavefront engineering to remove causes of artifacts in a hologram, according to an example.

[0053] FIG. 9 illustrates a method using multi-depths reconstruction for removing artifacts from a hologram, according to an example.

[0054] FIG. 10 shows a computer-implemented method for recognizing at least one characteristic of one or more objects contained in one or more droplets, according to an example.

[0055] FIG. 11 shows a computational model, such as a deep neural network, which can be used in the method of FIG. 10, according to an example.

[0056] FIG. 12 shows an apparatus for imaging and recognizing at least one characteristic of one or more objects contained in a droplet, according to an example.

[0057] FIG. 13 shows the in-line and lens-free holographic imaging configuration of the apparatus according to an example.

[0058] All the figures are schematic, not necessarily to scale, and generally only show parts which are necessary to elucidate example embodiments, wherein other parts may be omitted or merely suggested.

DETAILED DESCRIPTION

[0059] Example embodiments will now be described more fully hereinafter with reference to the accompanying drawings. That which is encompassed by the claims may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein; rather, these embodiments are provided by way of example. Furthermore, like numbers refer to the same or similar elements or components throughout.

[0060] FIG. 1 illustrates schematically a holographic imaging of an object 11 in a droplet 12. FIG. 1 shows on the top a physical shape of the droplet 12 without any object 11 inside, of an object 11 without the droplet 12, and of the droplet 12 including the object 11 inside. Then, in the middle of FIG. 1, the respective holograms after holographic imaging the physical shapes are illustrated. The droplet 12 alone results in a holographic signal 12a (left, e.g. referred to as a first ring pattern), the object 11 alone results in a holographic signal 11a (middle, e.g. referred to as a second ring pattern), and the combination of the droplet 12 and the object 11 in the droplet 12 results in a hologram 13 (right) based on a complex interaction of the holographic signal 11a and the holographic signal 12a (e.g., interactions of the first and second ring pattern). The holograms 13 can be reconstructed, wherein ideal reconstruction is assumed. The bottom of FIG. 1 shows images resulting from the reconstruction. The image 12b of only the droplet 12, the image 11b of only the object 11, and the combined image of the droplet 12 and object 11 are shown. Under ideal circumstances, these images resemble the physical shapes of the droplet 12 and object 11.

[0061] As has been discussed above, artifacts may appear in the hologram 13 of the object 11 contained in the droplet 12, which may make it hard to image and recognize characteristics of the object 11. Therefore, FIG. 2 shows a method 20 for imaging and recognizing at least one characteristic 25 of an object 11 contained in a droplet 12, wherein the method 20 can adjust for such artifacts.

[0062] The method 20 comprises a step 21 of generating at least one hologram 13 of the objects 11 contained in the droplet 12 using in-line and lens-free holographic imaging. For reasons of simplicity, like in FIG. 1, only one object 11 is shown in the droplet 12, however, the method 20 works with multiple objects 11 as well. The at least one hologram 13 may be generated using in-line lens-free imaging. As shown in FIG. 2, the at least one hologram 13 now includes at least one artifact 14. This at least one artifact 14 is caused by the presence of the droplet 12, and affects the at least one characteristic 25 of the object 11 contained in the droplet 12, which is to be recognized.

[0063] The method 20 further comprises a step 22 of removing fully or partially the at least one artifact 14, or the cause of the at least one artifact 14. Different ways to remove the artifact 14 or the cause of the artifact 14 are described below.

[0064] The method 20 further comprises a step 23 of generating an image 15 after or during removing the at least one artifact 14 or the cause of the at least one artifact 14. The image 15 may be generated by reconstruction of the hologram 13 after artifact removal, for example, by reconstruction methods known in the art. The image 15 comprises the object 11 and may comprise the droplet 12 as shown in FIG. 2. In another embodiment, the image 15 comprises the object 11 where the droplet 12 and the artifact are both removed.

[0065] The method 20 then comprises a step 24 of recognizing the at least one characteristic 25 of the object 11 based on the image 15, wherein the determination of the characteristic 25 is not anymore prevented or made difficult by the artifact 14. The at least one characteristic 25 may be directly derived from the image 15, for instance, if the characteristic is a shape, size, location, or the like. Other characteristics 25 could be obtained by post-processing the image 15.

[0066] FIG. 3 illustrates some possible causes of artifacts 14, which affect the object 11particularly make it difficult to derive the characteristic(s) 25and are caused by the presence of the droplet 12. For instance, an artifact 14 may be caused by an interference in the hologram 13 of the holographic signal 12a of the droplet 12 and the holographic signal 11a of the object 11 (e.g., interference of the first and second ring patterns). This is shown on the left side of FIG. 3, wherein the object signal is distorted by the droplet signal. This may also lead to an artifact in an image 31 after reconstruction of the hologram 13.

[0067] Also, noise in the holographic signal 12a of the droplet 12 and/or noise in the holographic signal 11a of the object 11 may cause an artifact 14. As shown in the middle of FIG. 3, an object signal loss may be caused with the noise from the droplet signal (e.g., the first ring pattern). This will result in loss of information of the object 11 in the image 31 as well. For example, ring-like artifacts 14 caused by the droplet 12 may come from the unknown phase of the hologram, which is due to the in-line lens-free holographic imaging. These artifacts 14 may occupy most of the area inside the droplet 12. For removing the artifact 14, the first ring pattern from the droplet 12 may thus be removed.

[0068] As shown on the right side of FIG. 3, typically there is a combination of many kinds of artifacts, including the previously described. In the following, some parameters are listed, which may cause such artifacts 14, and may be addressed by the method 20 of the present disclosure.

[0069] A size of the droplet 12.

[0070] A 3D shape of the droplet 12, e.g., round, square, an irregular shape, a contact angle difference, etc.

[0071] A mismatch of an optical property of the material of the droplet 12 and the environment.

[0072] A droplet density in case of multiple droplets 12.

[0073] An object density in a droplet 12, in case of multiple objects 11 inside a droplet 12.

[0074] A homogeneity of multiple droplets 12, in particular, a homogeneity in shape, size, and/or material.

[0075] A homogeneity of objects 11 inside a droplet 12, in particular, a homogeneity in size, shape, and/or type of object 11.

[0076] Foreign objects (not to be analyzed) in the droplet 12 (e.g., inside, outside, variations of the foreign objects in shape, size, materials, etc.).

[0077] Multiple droplets 12 at different heights/depth into the imaging direction.

[0078] Multiple objects 11 located at different heights/depth inside a droplet 12 in the imaging direction.

[0079] FIG. 4 shows a pipeline, also referred to as meta-pipeline, for removing at least one artifact 14 or the cause of the at least one artifact 14 from a hologram 13. The pipeline is applicable for both single and multi-wavelength in-line lens-free holographic imaging. Notably, multi-aperture can be also considered to be a special case of multi-wavelengths.

[0080] The pipeline may provide various methods for removing the at least one artifact 14 or the cause of the at least one artifact 14. In particular, the pipeline may offer wavefront engineering 41, deep learning filtering 42x, a FISTA algorithm or a similar algorithm, conventional image and signal processing 44, and multi-depths reconstruction 46 as possible methods to remove the artifact 14.

[0081] One or more of the various methods may be selected, for instance, depending on the type of the at least one artifact 14 and/or based on environmental conditions. In the example of FIG. 4, the hologram 13 is input into the pipeline. At first, the pipeline determines, whether a lensing effect of the droplet 12 causes information loss, for example, at the edge or of local saturation. If this is the case, then wavefront engineering 41 may be selected.

[0082] If this is not the case, the pipeline further determines whether the at least one artifact 14 and the object 11 are consistent. If yes, then deep learning 42x can be applied. Afterwards, any reconstruction method 45e.g. as described in the literaturecan be applied to reconstruct the image 15 from the hologram 13.

[0083] After applying the wavefront engineering 41, the pipeline may further determine how the artifact 14 caused by the droplet 12 is compared to the object signal. Possibilities are strong or not regular, weak or regular, or no artifacts. If there are no artifacts 14, the conventional reconstruction 45 can be used as described above.

[0084] In the case of stronger or not regular artifacts 14, the FISTA algorithm 43 can be applied. In the case of weaker or regular artifacts 14, conventional image and signal processing 44 may be applied, and may be sufficient to remove such artifacts 14.

[0085] Subsequently, the pipeline may further determine whether there are objects 11 or droplets 12 at different heights/depth of the imaging direction. If yes, then multi-depths reconstruction 46 can be applied, which may be followed by post-processing afterwards, to generate the final image 15. If not, then the post-processing can be applied directly.

[0086] FIG. 5 compares various methods of removing the at least one artifact 14 or the cause of the at least one artifact 14 from the hologram 13, as shown in FIG. 4. In particular, FIG. 5 shows a trade-off between a strength of improvement of the respective method (on the y-axis) and effort, e.g., computational or physical, of the respective method (on the x-axis).

[0087] In view of FIG. 5, the method 20 may further include a step of determining a strength of the at least one artifact 14, and a step of selecting how to remove the at least one artifact 14 or the cause of the at least one artifact 14 based on the determined strength. The strength of the artifact 14 or cause may thereby be set in correspondence with the improvement shown in FIG. 5. That is, the stronger (higher strength) the artifact 14 or cause is, the larger the improvement achieved with the selected method can be.

[0088] Notably, the strength of the at least one artifact 14 may be determined by comparing at least one parameter of the object 11 in the hologram 13 (with the droplet 12) and in at least one reference hologram of the object 11 without droplet 12.

[0089] In the following, the various methods for removing the artifact(s) 14 or cause of the artifact(s) 14 are illustrated and explained.

[0090] FIG. 6 shows a method based on using the FISTA algorithm 43 for removing stronger artifacts 14 from the hologram 13. The method 20 comprises, in this case, a step of processing the at least one hologram 13 with one or more numerical optimization algorithms. These algorithms may include the FISTA algorithm 43 as shown. Also, an ADMM algorithm may be included instead of FISTA. The FISTA algorithm 43 may be performed based on certain constraints (e.g., 11 constraint to minimize a ring effect from the droplet and/or a box constraint to remove some bright artifact), which may be selected depending on the determined artifact(s) 14 or the desired characteristics 25 to be recognized. The FISTA algorithm 43 may be performed iteratively, to remove unwanted signals and artifacts 14 of different kinds form the hologram 13. The FISTA algorithm 43 itself is known, see e.g., Fast compressive lens-free tomography for 3D biological cell culture imaging Luo, Zhenxiang, et al., Optics Express 28.18 (2020): 26935-26952.

[0091] Before applying the FISTA algorithm 43, regular pre-processing steps may be applied to the hologram 13, for instance illumination balancing. After applying the FISTA algorithm 43, a high-quality reconstruction may be performed to obtain the image 15.

[0092] FIG. 7 shows conventional image and signal processing for removing weaker artifacts 14 from the hologram 13. The method 20 comprises, in this case, removing 22 the at least one artifact 14 or the cause of the at least one artifact 14 by processing the at least one hologram 13 by at least one image processing method 44. The image processing method 44 may comprise denoising, local illumination balancing and enhancement, using filters to remove unwanted signals, or wavelet transfer etc. After the image processing method 44 is applied, a reconstruction may be made of the hologram 13 to obtain the image 15. Afterwards, conventional image processing methods may be applied to the image 15. These may include signal or image enhancement, filters to remove unwanted signals, or image engineering to detect artifacts and remove them.

[0093] FIG. 8a, FIG. 8b, and FIG. 8c illustrate wavefront engineering 41 for removing causes of artifacts 14. The method 20 comprises, in this case, a step of regenerating the hologram 13 of the objects 11 contained in the droplet 12 using in-line lens-free holographic imaging, wherein a second illumination wavefront is used for regenerating the hologram 13. The second illumination wavefront is different from a first illumination wavefront that was used for generating 23 the hologram 13. The second illumination wavefront comprises a single wavelength or multiple wavelengths.

[0094] FIG. 8a shows examples in the case of a normal illumination during the in-line lens-free holographic imaging. In particular, FIG. 8a shows on the left side an example of a droplet 12 of a non-polar material in a medium of a polar material, e.g., an oil droplet 12 in water. Further, FIG. 8a shows on the right side an example of a droplet 12 of a polar material in a medium of a non-polar material, e.g., a water droplet 12 in oil. Lensing effects of the droplet 12 are illustrated by the arrows after the droplet 12. FIG. 8b shows the same examples in case of an engineered illumination, which may use an illumination wavefront specifically selected to avoid the lensing effects of the droplet 12, as illustrated by the more parallel arrows after the droplet 12. These lensing effects are due to the different materials (e.g. oil and water). However, as shown in FIG. 8c, lensing effects are also a concern for certain droplet shapes, e.g., for flat droplets 12.

[0095] FIG. 9 illustrates a method using multi-depths reconstruction for removing artifacts 14 from the hologram 13. The method 10 comprises, in this case, as step of generating the image 15 by multi-depth reconstruction 46 of the at least one hologram 13 obtained by in-line lens-free holographic imaging. That means the reconstruction of the hologram 13 is performed at multiple depths (in direction of the imaging, i.e., illumination wavefront), so that a projection onto the planar image 15 is achieved. Notably, the multi-depths reconstruction can also be applied to single droplet/object levels.

[0096] FIG. 10 shows a flow-diagram of a computer-implemented method 100. The method 100 is for recognizing at least one characteristic 25 of one or more objects 11 contained in a droplet 12. The method 100 is in part similar to the method 20.

[0097] The method 100 comprises a step 101 of obtaining at least one hologram 13 of the one or more objects 11 contained in the droplet 12. For instance, the at least one hologram 13 may be provided to the computer or processor that implements the method 100, and/or may be obtained from a holographic imager. The at least one hologram 13 has been previously generated using in-line lens-free imaging, for example, by the holographic imager. As in the method 20 of FIG. 2, the at least one hologram 13 includes at least one artifact 14, which is at least caused by the presence of the droplet 12 and affects the at least one characteristic 25 of the one or more objects 11 contained in the droplet 12.

[0098] The method 100 further comprises a step 102 of processing the at least one hologram 13 using a computational model 42, for instance, a neural network. The computational model 42 removes fully or partially the at least one artifact 14. The hologram 13 with the artifact(s) 14 may be input into the computational model 42, and the hologram without the artifact(s) may be generated as output by the computational model 42. The computational model 42 may be a trained and/or trainable model.

[0099] The method 100 further comprises a step 103 of generating an image 15, after or during the removing step 102 of the at least one artifact 14, wherein the image 15 comprises the one or more objects 11. The image 15 may be generated by reconstruction of the hologram 13. Then, the method 100 comprises a step 104 of recognizing the at least one characteristic 25 of the one or more objects 11 based on the image 15. The steps 103 and 104 may be either or both performed by the computational model 42.

[0100] According to the above, the computational model 42 may remove artifacts 14 caused by droplets 12 on the hologram level, prior to reconstruction. FIG. 11 shows a computational model, particularly a deep neural network, which can be used for the method of FIG. 10. The shown computational model 42 is, for example, a 4-level UNet with contraction blocks, expansion blocks, and residual blocks.

[0101] A simulated dataset may be used to train the computational model 42, wherein the following parameters may generally be considered in the simulated dataset: experimental conditions; desired artifacts properties; or expected objects properties. Specifically, the dataset may include or be based on the certain parameters e.g. sensor specifications, optical parameters of the setup, the morphological features (like size, shape, etc.) and properties (for example, reflex index, which is related to the materials) of the droplets, objects, properties of the artifacts, or simulated setup configurations.

[0102] During the training of the computational model 42, the input may be a signal that includes both a plurality of possible artifacts 14 and a plurality of objects 11 in any combination with another. The expected output to train the model 42 is a signal comprising the objects 11 alone. During the training, a loss may be a mean squared error (MSE) loss, an ADAM optimizer may be used, a starting learning rate (LR) may be small, and an LR update policy may be employed.

[0103] The main observations of using the computation model 42 are that the result is quite satisfactory on the simulated data, and that the removal of the artifacts 14 from holograms 13 including droplet signal and the object signal works well.

[0104] FIG. 12 shows schematically an apparatus 1200, for imaging and recognizing at least one characteristic 25 of one or more objects 11 contained in a droplet 12. The apparatus 12x may perform the method 20 of FIG. 2.

[0105] The apparatus 1200, to this end, comprises a holographic imager 1201, which is configured to generate at least one hologram 13 of the one or more objects 11 contained in the droplet 12 using in-line lens-free holographic imaging. The holographic imager 1201 may to this end illuminate the droplet 12 including the object 11, for instance, with a selected wavefront including one or more wavelengths. Wavefront engineering 41 may be applied using the holographic imager 1201. Like in FIG. 2, the at least one hologram 13 includes at least one artifact 14, which is at least caused by the presence of the droplet 12, and affects the at least one characteristic 25 of the one or more objects 11 in the droplet 12.

[0106] The apparatus 1200 further comprises a processor 1202, which is configured to remove fully or partially the at least one artifact 14 or the cause of the at least one artifact 14. The processor 1202 is further configured to generate an image 15, after or during removing the at least one artifact 14 or the cause of the at least one artifact 14, comprising the one or more objects 11, and to recognize the at least one characteristic 25 of the one or more objects 11 based on the image 15. The processor 1202 is accordingly configured to perform artifact removal and reconstruction of the hologram 13, for example, by reconstruction methods known in the art.

[0107] The processor 1202 may comprise processing circuitry (not shown) configured to perform, conduct, or initiate the various operations described. The processing circuitry may comprise hardware and/or the processing circuitry may be controlled by software. The hardware may comprise analog circuitry or digital circuitry, or both analog and digital circuitry. The digital circuitry may comprise components such as application-specific integrated circuits (ASICs), field-programmable arrays (FPGAs), digital signal processors (DSPs), or multi-purpose processors. The processor 1202 may further comprise memory circuitry, which stores one or more instruction(s) that can be executed by the processor 1202 or by the processing circuitry, in particular under control of the software. For instance, the memory circuitry may comprise a non-transitory storage medium storing executable software code which, when executed by the processor or the processing circuitry, causes the various operations of the processor 1202 to be performed.

[0108] FIG. 13 shows that the apparatus 1200 performs in-line lens-free holographic imaging. The holographic imager 1201 of the apparatus 1200 comprises a light source 1201a. With this light source 1201a, the holographic imager 1201 is configured to illuminate the droplet 12 including the at least one object 11. For reasons of simplicity, again only one object 11 is shown in the droplet 12. Also, only one droplet 12 is shown, however, the apparatus 1200 works likewise with multiple objects 11 and/or multiple droplets 12. The light waves emitted by the light source 1021a may partly interact with the droplet 12, particularly with the object 11, and may partly go through the droplet 12 without interaction.

[0109] The holographic imager 1201 further comprises a recording medium 1201b, typically a digital sensor, which is configured to record an interference pattern of the respective light wavesi.e., those light waves that have interacted with the droplet/object and those light waves which have not. The holographic imager 1201 is in this way configured to obtain the hologram 13. Since the recording medium 1201b is arranged behind the droplet/object along the light path originating at the light source 1201a, and is thus arranged to record also directly light waves passing through the droplet 12, the holographic imager 1201 has an in-line configuration for in-line holographic imaging. Moreover, no lens is needed in this configuration, because no light beam needs to be deflected, so the holographic imaging performed by the apparatus 1200 is also lens-free. Generally, in-line lens-free holographic imaging captures holograms 13 without using lenses by recording interference patterns directly on the recording medium 1201b, wherein a coherent light source 1201a aligned with the droplet 12 and recording medium 1201b may be used as an example (but not required).

[0110] The processor 1202 of the apparatus 1200 is further able perform the reconstruction of the hologram 13 after or during artifact removal, for example, by reconstruction methods known in the art, to obtain the image 15.

[0111] In sum, the present disclosure provides a method for using in-line lens-free holographic imaging to obtain high-quality images 15 of objects 11 in droplets 12, and to obtain characteristics 25 of the objects 11 based on the images 15.

[0112] In the claims as well as in the description of this disclosure, the word comprising does not exclude other elements or steps and the indefinite article a or an does not exclude a plurality. A single element may fulfill the functions of several entities or items recited in the claims. The mere fact that certain measures are recited in different dependent claims does not indicate that a combination of these measures cannot be used.

[0113] While some embodiments have been illustrated and described in detail in the appended drawings and the foregoing description, such illustration and description are to be considered illustrative and not restrictive. Other variations to the disclosed embodiments can be understood and effected in practicing the claims, from a study of the drawings, the disclosure, and the appended claims. The mere fact that certain measures or features are recited in mutually different dependent claims does not indicate that a combination of these measures or features cannot be used. Any reference signs in the claims should not be construed as limiting the scope.