JP5869920B2 - Image evaluation method - Google Patents

Description

  The present invention relates to a technique for processing an image obtained by imaging a recess provided in a sample holding plate and capable of holding a liquid, and more particularly to a technique for evaluating the significance of the image.

  In a medical or biological science experiment, for example, a liquid or gel-like fluid flows into each well of a plate-like instrument (for example, called a microplate or a microtiter plate) provided with a large number of depressions called wells. A body (for example, a culture solution, a medium, etc.) is injected, and the cells cultured here are observed and measured as a sample. In recent years, a sample is imaged with a CCD camera or the like and converted into data, and various image analysis techniques are applied to the image data for observation and analysis. In such imaging, unevenness in brightness (shading) may occur in the obtained image due to the non-uniform amount of illumination light incident on the sample.

  As a technique for eliminating such unevenness, it is conceivable to apply a shading correction technique put into practical use in a digital camera or the like. For example, Patent Document 1 describes a shading correction technique that can cope with a case where the shading characteristic is asymmetric with respect to the optical axis or a case where there is variation among individuals.

Japanese Patent Laying-Open No. 2008-177794 (for example, FIG. 3)

  In addition to the above, in the image of the recess into which the liquid has been injected, image unevenness due to refraction at the liquid surface may occur. In particular, when imaging is performed by applying illumination light from above the recess and receiving light transmitted downward, the shadow of the wall surface of the recess appears in the image, and the brightness of the peripheral edge of the recess decreases. In this part, it may interfere with the observation of the sample. The degree of shadow reflection varies depending on the state of the liquid surface, and the state of the liquid surface varies from sample to sample depending on the amount of liquid, the wettability with respect to the wall surface of the recess, the working condition at the time of injection, and the like.

  On the other hand, in the conventional shading correction techniques including those described in Patent Document 1 described above, correction is performed based on a comparison with the light quantity distribution measured in advance, so that the light quantity distribution itself varies from sample to sample. There is a possibility that a sufficient correction effect cannot be obtained with respect to an image obtained by capturing the image.

  Therefore, it is necessary to establish a shading correction technique that can be applied to an image obtained by imaging such an object, or an imaging technique (including illumination technique) that can eliminate the influence of shadows. It becomes. However, the technology itself for objectively and quantitatively evaluating the influence of the shadow reflected in the image has not yet been established.

  This invention is made | formed in view of the said subject, and it aims at providing the technique which can evaluate appropriately the significance of the image which imaged the hollow part in which the liquid was inject | poured.

In order to achieve the above object, the image evaluation method according to the present invention acquires an image data acquisition step of acquiring image data obtained by imaging the recess in a state where a liquid is injected into the recess provided in the sample holding plate. Setting a straight line across the recess in the image of the recess, a profile deriving step for obtaining an image density profile in the recess along the straight line based on the image data, and determining a gradient of the image density profile, A determination step of comparing the value with a predetermined threshold value to determine the significance of the image, and in the determination step, the image is significant when the maximum absolute value of the gradient is smaller than the threshold value. It is determined that the image is not significant when it is greater than the threshold .

  In the present invention, the “liquid” includes a paste or gel, and “the state in which the liquid is injected into the recess” refers to the state in which the injected liquid is held in the recess as it is, A state in which a liquid having a low viscosity immediately after injection is cured in a solid or gel state is also included.

  In the invention configured as described above, the significance of the image is determined based on the image density profile along a straight line crossing the depression in the image. When the effect of the shadow on the wall surface of the depression appears significantly in the captured image, the background density of the image changes abruptly at the boundary between the portion where the shadow is reflected and the portion where the shadow is not. Therefore, in the present invention, the gradient of the image density profile on the straight line crossing the depression is obtained, and the significance of the image is judged by comparing the value with a predetermined threshold value. By appropriately setting the threshold value according to the purpose, it is possible to appropriately evaluate an image in consideration of the density change caused by the shadow of the depression wall surface.

  If shadows are reflected in the entire image of the depression, the background density of the entire depression will be affected, but this effect will also appear on the observation object in the depression. The influence on itself is small, and correction for removing the influence is relatively easy. On the other hand, when a shadow is partially reflected in the image of the depression, the background density of the observation object varies and correction is more difficult. The latter shadow is therefore a more serious problem. The present invention is suitable for finding an image in which the influence of such a shadow is expected.

Specifically, in the judgment step, the image is judged that there is significant when the maximum value of the absolute value of the slope is less than the threshold value, you determined that the image has no significance when greater than the threshold value . When the maximum value of the gradient (absolute value) is a large value, it means that there is a rapid concentration change on the straight line. Therefore, by making an image having such a large density change insignificant, it becomes possible to distinguish an image that is significantly affected by the reflection of a shadow from an image that is less affected.

  Further, for example, in the profile deriving step, an image density profile may be derived by performing a smoothing process on a data sequence in which pixel values of pixels on a straight line are arranged in the arrangement order. The shadow image on the wall surface of the depression appears in a continuous region in the image, but may include a local concentration change due to, for example, foreign matter mixed in the liquid or bubbles on the liquid surface. By performing the smoothing process on the pixel value data string to obtain the image density profile, it is possible to avoid such a local density change from affecting the determination result.

  Further, for example, in the determination step, a difference between adjacent sampling data in a data string obtained by sampling pixel values of pixels on a straight line in an arrangement order may be used as a gradient value. Good. In this case, sampling may be performed on the data string subjected to the smoothing process described above. In general, it is considered that the shadow of the wall surface of the depression has a sufficiently large size with respect to the pixel size in the imaging for the contents of the depression. Therefore, it is possible to make a sufficiently accurate determination based on the sampled data string, and it is possible to make the influence of the local concentration change as described above less susceptible.

  Further, for example, in the profile derivation step, a plurality of different straight lines are set and an image density profile is obtained for each of the straight lines, while in the determination step, the gradient of the image density profile along the straight line is determined among the plurality of straight lines. If there is at least one whose maximum value exceeds the threshold, it may be determined that the image is not significant. The surface of the liquid injected into the depression may have irregular irregularities, and as a result, the shadow of the depression wall surface also appears irregularly in the image. Such irregular shadows can be evaluated by obtaining image density profiles for a plurality of different straight lines. In addition, when there is at least one of the gradients whose maximum value exceeds the threshold, it is determined that the image is not significant, and the influence of the shadow that appears only at a specific position is overlooked. Is prevented.

  In this case, for example, a plurality of straight lines that intersect with each other in the recess may be set. By doing so, the significance is determined based on the image density profiles obtained for various directions in the image. Therefore, when the shadow has anisotropy (that is, asymmetric) in the image. Can also respond.

  More specifically, for example, when the upper surface opening of the recess is substantially circular, a plurality of straight lines may be set to pass through the center of the recess. In such a depression, the shadow of the wall surface appears at the peripheral edge of the image of the depression. By determining the image density profile in the radial direction from the center of the recess and in various directions, it is possible to appropriately evaluate the influence of such a shadow.

  Further, for example, in the image data acquisition step, image data corresponding to an image including a plurality of depressions provided on the sample holding plate is acquired, and the profile derivation step and the determination step are executed for each of the plurality of depressions. Also good. When the sample holding plate is provided with a plurality of depressions, the light quantity distribution of light incident on each depression varies, and the surface state of the injected liquid also varies from depression to depression. Therefore, it is more preferable to evaluate each recess.

  The present invention exhibits a particularly excellent effect by being applied to the evaluation of an image obtained by irradiating light from above the sample holding plate and receiving light transmitted through the lower part of the sample. It is. As described above, when such an imaging method is adopted, shadow reflection due to refraction of irradiation light on the liquid surface is remarkable. Therefore, by applying the present invention to the evaluation of an image captured in such a manner, information serving as a criterion for determining the validity of the imaging method itself can be obtained. Such information is useful for improving the imaging method including the illumination method.

  According to the present invention, since the significance is determined based on the gradient of the density profile of the image obtained by imaging the depression into which the liquid is injected, the quality of the image in which the shadow of the depression wall is reflected is objectively determined. And can be quantitatively evaluated.

It is a figure which shows schematic structure of the imaging device which can apply this invention. It is a figure which shows the more detailed structure of an imaging part. It is a figure which shows the example of the image imaged with the imaging device of FIG. It is the figure which showed the well image typically. It is a flowchart which shows one Embodiment of the image evaluation method concerning this invention. It is a figure which shows the example of a setting of the line segment in this embodiment. It is a figure which shows the example of the image used for evaluation.

  FIG. 1 is a diagram showing a schematic configuration of an imaging apparatus to which the present invention can be applied. As shown in FIG. 1A, the imaging apparatus 1 includes a plurality of, for example, 96 (12 × 8), each of which is filled with a liquid such as a culture solution, a medium, or a reagent (only a part of which is shown). A holder 11 that holds the microplate M in a substantially horizontal state by abutting against the peripheral edge of the lower surface of the sample (microplate) M in which the well W of the matrix arrangement is formed, and a light source 12 provided above the holder 11 And an imaging unit 13 provided below the holder 11 and a control unit 10 that controls these to execute a predetermined operation. For the following explanation, coordinate axes are set as shown in FIG. The XY plane is a horizontal plane, and the Z axis is a vertical axis.

  The diameter and depth of each well W in the microplate M are typically about several mm. As an example, the dimension of each part of the microplate M used in the experiment described later is shown. As shown in FIG. 1B, the diameter Dt of the opening Wt at the top of each well W is 6.69 mm, while the inner diameter Db of the well bottom surface Wb is 6.58 mm. As can be seen, the inner wall surface Ws of the well W is not a simple cylindrical surface, but has a tapered shape with the side surfaces inclined obliquely. The depth Dd of the wells W is 10.9 mm, and the arrangement pitch Dp of the plurality of wells W is 9 mm. In addition, the dimension of each part is a mere example, Comprising: The size of the microplate which this imaging device 1 makes object is not limited to these, It is arbitrary.

  The light source 12 is controlled by a light source control unit 112 provided in the control unit 10 and is collectively applied to a plurality of wells W from above the microplate M held by the holder 11 in accordance with a control command from the light source control unit 112. Then, the light L is irradiated. The irradiated light is visible light, and white light is particularly preferable.

  The imaging unit 13 functions as a camera that captures an image of the microplate M by receiving transmitted light Lt emitted from the light source 12 and transmitted below the microplate M held by the holder 11. . The imaging unit 13 is connected to a camera driving mechanism 113 provided in the control unit 10, and the camera driving mechanism 113 moves the imaging unit 13 along a horizontal plane (XY) along the lower surface of the microplate M held by the holder 11. Scanning movement within a plane).

  That is, in this embodiment, the imaging unit 13 can be scanned and moved along the lower surface of the microplate M. Here, the imaging unit 13 moves relative to the microplate M. However, it is sufficient that the relative movement between the imaging unit 13 and the microplate M is realized. In this sense, the microplate M is moved relative to the imaging unit 13. You may make it move.

  Image data picked up by the image pickup unit 13 is given to the image processing unit 114. The image processing unit 114 appropriately performs image processing on the image data from the imaging unit 13 or executes predetermined calculation processing based on the image data. Data before and after processing is stored and saved in the storage unit 115 as necessary.

  The imaging device 1 is configured such that a liquid or the like held in each well W (in this specification, a liquid, a gel-like or semi-fluid solid, and a well in a fluid state such as soft agar) It is a generic name for those that are injected and then solidified), and more specifically, an optical image (bright field image) of an imaging object such as a cell or the like contained therein, or having a predetermined optical characteristic from the optical image The present invention can be applied to the use of detecting a specific portion having an optical characteristic different from that of the liquid or the like held in the well W by utilizing the difference in the optical characteristic. For example, it can be suitably used for the purpose of imaging cells and cell clumps (spheroids) in a culture medium as an imaging target, and further automatically detecting such cells and the like by image processing.

  FIG. 2 is a diagram illustrating a more detailed configuration of the imaging unit. As shown in FIG. 2A, the imaging unit 13 outputs light emitted from the bottom surface of the microplate M held by the holder 11 and the line sensor 131 that outputs an electrical signal corresponding to incident light, for example, a CCD. And an imaging optical system 132 that forms an image on the light receiving surface of the line sensor 131. The imaging optical system 132 may include a plurality of optical components such as lenses, but here, for easy understanding, a single lens is representatively shown here.

  The line sensor 131 is a one-dimensional array of a large number of fine imaging elements 131a in the Y direction. In the longitudinal direction, the entire at least one well W, more preferably a plurality of (the same In the drawing, three wells W are configured to be included in the imaging range SR at a time. In the figure, the length of the line sensor 131 in the Y direction is indicated by a symbol w, and the length of the visual field on the bottom surface of the microplate M is indicated by a symbol w ′.

  As shown in FIG. 2B, the scanning movement direction of the line sensor 131 by the camera driving mechanism 113 is the X direction. In this way, the line sensor 131 in which the image pickup devices are arranged along the Y direction is scanned and moved in the X direction along the bottom surface of the microplate M, thereby capturing a two-dimensional image of the microplate M viewed from the bottom surface side. Is possible. In addition, by repeating the scanning movement with the line sensor 131 at different positions in the Y direction, a large number of wells W formed on the microplate M can be sequentially imaged.

  The line sensor 131 can obtain a high-definition image because the pixel size of each image sensor is small. In addition, an imaging optical system 132 is configured so that a large number of imaging elements are arranged in a line and an optical image of each part of the well W is formed on each imaging element, and these are arranged at appropriate positions. By making more light from the well W enter the line sensor 131, the time required to image one well W is shortened. Thereby, imaging about many wells W can be performed at high speed.

  Consider a case where the image evaluation method according to the present invention is applied to the imaging apparatus 1 configured as described above. The image evaluation method according to the present invention evaluates the significance of a captured image, and is particularly suitable for evaluating an image captured through a liquid surface. When this is applied to the imaging apparatus 1 described above, first, it can be used for evaluation of an image obtained by imaging a biological sample such as a cell cultured in a medium injected into the well W as described above. . In this case, the captured image is analyzed by various image processing, and the shape and distribution of cells are detected. However, it is necessary to evaluate whether the original image can withstand such image processing. It is. That is, there is a possibility that the image is not worthy of analysis due to the reflection of a shadow at the time of imaging.

  FIG. 3 is a diagram illustrating an example of an image captured by the imaging apparatus of FIG. FIG. 3A is an example of an image of a well W including a culture medium and cell clumps cultured in the medium, which is imaged under favorable illumination conditions. In this example, the central portion of the circular region corresponding to the well is brightest, and the peripheral portion of the circular region is slightly darker than this. This is considered to be due to the refraction and scattering of the illumination light on the liquid surface, but it is still possible to clearly distinguish the cell clumps located at the peripheral edge.

  On the other hand, FIG. 3B is an example of an image obtained by imaging the same well W under illumination conditions different from the above, and the peripheral portion becomes darker, and it is difficult to distinguish the cell cluster located in this portion. It has become. Such darkness of the peripheral portion and its spread show different aspects depending on the illumination conditions, the amount of liquid injected into the well W, and the surface state thereof.

  FIG. 4 is a diagram schematically showing a well image. FIG. 4A is a schematic diagram of the image of FIG. 3B, and FIG. 4B shows the luminance value of each pixel located on the line segment QQ ′ shown in FIG. This is a plotted graph. The well image Iw has a bright part area BR that is relatively bright (high brightness) in the center part of the substantially circular well area WR corresponding to the well W, and has a dark part dark part DR in the peripheral part. . The bright area BR is not necessarily a circle that is coaxial with the circle corresponding to the contour of the well W, but has a distorted shape.

  It is considered that the dark region DR is generated in the well image Iw due to the refraction / scattering of illumination light on the liquid surface and the shadow on the well wall surface. In particular, it is considered that the shadow of the inner wall surface of the well W exposed above the liquid level is reflected in the image due to refraction at the liquid level. In particular, when the amount of liquid is small or when the well wall surface is made of an opaque material, a dark region DR having a high concentration (low luminance) appears.

  The following can be considered as the influence caused by the occurrence of such a dark region DR. First, among the cells or cell clumps Sp distributed in the well W, cell clumps (for example, indicated by the symbol Sp1) existing in the bright region BR are detected and analyzed well without being affected by the background. Is possible. On the other hand, for the cell agglomerates (for example, indicated by the symbol Sp2) present in the dark region DR, the detection itself may be impossible due to the dark background, or a large error may be caused in the detection of the concentration. In addition, for a cell clump located near the boundary between the bright area BR and the dark area DR (for example, indicated by reference numeral Sp3), the outline of the cell clump becomes unclear, which causes erroneous detection of size and concentration. obtain.

  Therefore, it is desirable to exclude an image from which an error due to the dark region DR is expected as described above from being processed as insignificant, or at least notify the operator of that fact. The present invention can be suitably applied to such image evaluation.

  Further, as a second case in which the image evaluation method of the present invention is applied to the imaging apparatus 1, it can be applied to the evaluation of the imaging apparatus itself in the development or manufacturing process of the imaging apparatus 1. As described above, the state of the dark area DR generated in the image due to the shadow on the well wall changes greatly depending on the imaging condition including the illumination condition, and the influence of the shadow can be reduced by devising the imaging condition. In such applications, the image evaluation method of the present invention can be suitably applied as a reference for determining whether the imaging conditions are good or bad. In addition, as a product inspection in the manufacturing process of the imaging device, an application of providing a quality check process to which the image evaluation method according to the present invention is applied is also conceivable.

  In particular, in an apparatus that comprehensively images a plurality of wells W provided on the microplate M, such as the imaging apparatus 1, the influence of shadows varies depending on the illumination conditions for each well. Therefore, while it is necessary to evaluate the imaging conditions for each well individually, it is not practical for the operator to visually check them one by one, and the images of each well are objectively and quantitatively evaluated. It is very meaningful if possible.

  Specifically, the microplate M in which an appropriate liquid (or liquid or the like) is injected into each well W is set in the imaging apparatus 1 and imaged in the same manner as the above-described imaging of the sample. By evaluating the image of each well W thus imaged and outputting the result, it is possible to effectively assist the operator in determining whether the imaging conditions are appropriate for the apparatus. A specific operation for enabling this will be described next.

  FIG. 5 is a flowchart showing an embodiment of an image evaluation method according to the present invention. This process is executed by the control unit 10 executing a predetermined control program to control each unit of the imaging apparatus 1. First, as described above, the microplate M in which an appropriate liquid or the like is injected into each well W is set in the imaging device 1 to perform imaging, and an original image is acquired (step S101). Here, the liquid injected into the well W is a dummy for evaluation, and an appropriate one can be used. However, a liquid close to a medium used for an actual sample is desirable in terms of viscosity and optical characteristics. Further, the liquid injection may be performed in another place before the microplate M is set in the imaging apparatus 1. At this time, it is desirable that the microplate M be held in a horizontal posture. In particular, in the case where the liquid is cured by injection, the microplate M is kept horizontal until at least the injected liquid is cured. It is desirable to stand in a posture. Note that the microplate M in which the injected liquid is hardened does not necessarily have to be in a horizontal posture at the time of imaging.

  Next, only the region corresponding to each well W is cut out as a well image Iw from the image of the entire plate thus imaged (step S102). Subsequent processing is performed for each well image. In the following, a process for one of the well images thus cut out will be described, but the same process is applied to all the well images. Subsequently, a plurality of line segments that cross the well W are set for one well image (step S103), and pixel values of each pixel in the well located on each line segment are extracted (step S104). ).

  FIG. 6 is a diagram showing an example of setting a line segment in this embodiment. As shown in FIG. 6A, here, four line segments A-A ', B-B', C-C ', and D-D' are set for one well image Iw1. Each line segment is set at equiangular intervals so as to cross each other in the vicinity of the center of the well image Iw1. Therefore, in this example, the angle formed by the adjacent line segment is 45 degrees.

  FIG. 6B schematically shows a luminance profile obtained from a data string of pixel values (luminance) of each pixel on one line segment (for example, line segment A-A ′). As indicated by a broken line, the luminance of each pixel is high at a position corresponding to the bright area BR1 in the central portion of the well, and the luminance of each pixel is low at a position corresponding to the dark area DR1 in the peripheral portion. The luminance changes greatly in the vicinity of the boundary between the bright region BR1 and the dark region DR1, and when the absolute value of the gradient (differential value) of the curve representing the luminance is plotted, as shown by the solid line in the figure, the well region WR1 A relatively high value is shown in the portion corresponding to the outline of the light source region and the vicinity of the boundary between the bright region BR1 and the dark region DR1, and a relatively low value is shown in the other portions.

  Here, when image processing is performed on an image including a dark area, a sharp luminance change at the boundary between the dark area and the bright area becomes more problematic than the brightness (or darkness) of the entire image. There are many cases. That is, in an image of a sample containing cells or the like, even if the brightness of the entire image is insufficient, if the luminance change of the background portion of cells or the like is gentle, for example, a background image is created and It is relatively easy to reduce the influence by appropriate image processing such as taking a difference. On the other hand, when the brightness change in the background portion is steep at the boundary between the dark area and the bright area, cells in the vicinity of the boundary are regarded as the background and disappear from the processed image, or the shape is There are problems such as being detected as different from the original one, and it is more difficult to eliminate the influence.

  Therefore, in this embodiment, a predetermined threshold Gth is preset for the luminance gradient (more precisely, its absolute value), and if the luminance gradient exceeds this threshold Gth, the image is significant. Judge that there is no. By doing so, it is previously grasped that a dark area may be formed in the image, which may affect subsequent image analysis. Such information is useful, for example, for improving imaging conditions including an illumination light source, managing manufacturing quality, and developing new image processing techniques in shading correction and the like.

  Note that the appearance of the dark area in the well image is not necessarily symmetric as described above, and has anisotropy as shown in FIG. Therefore, by setting a plurality of line segments in different directions and making the above determination on each line segment, it is possible to avoid overlooking the influence of the shadow appearing in a specific direction. Here, four line segments are set at equiangular intervals, but the number and arrangement of the line segments are not limited to this and are arbitrary. However, since the shadow of the well wall surface appears at the peripheral edge of the well image, it is effective to obtain the luminance profile and its gradient on a line segment set radially from the center of the well. In particular, when the well cross-sectional shape is circular, it is desirable to set a line segment that intersects at the center of the circle.

  Returning to FIG. 5, the description of the flowchart will be continued. In principle, it is only necessary to determine the significance of the image based on the gradient of the luminance profile on the line segment set as described above. However, in the actual image, the image was mixed with the liquid injected into the well. The brightness profile may be disturbed due to foreign matters or bubbles. As shown in FIG. 3B or FIG. 6A, the shadow of the well wall surface is reflected so as to continuously cover the peripheral edge of the well, and it does not have a fine structure even when considered from the principle.

  If the resolution of the imaging unit 13 is about 2400 dpi (dots per inch), for example, the size of one pixel in the image is about 10 μm. On the other hand, the size of the cell clump which is the object to be imaged is generally about 40 μm to 50 μm. Therefore, in order to prevent erroneous detection of cell clumps due to abrupt changes in background density, it is important to know density changes in dimensions of several pixels.

  From these points, regarding the luminance profile, information on detailed changes in units of pixels is not always necessary, but rather there is a concern about the influence of the above-described contamination with foreign substances. Therefore, the smoothing process (step S105) and the sampling process (step S106) are performed on the pixel value data string for each pixel. As the smoothing process, for example, a moving average process between successive pixels can be applied, and the calculation range can be, for example, 7 pixels centered on the pixel. Further, the sampling interval of the sampling process can be about 4 to 5 pixels at the maximum due to the above-described size relationship. These smoothing processing and sampling processing are not essential, and may be omitted depending on the state of the obtained data.

From the data string of the processed data V (i) (where i = 1, 2,...) Obtained in this way, the absolute value of the difference between adjacent data:
D (i) = | V (i + 1) −V (i) | (i = 1, 2,...)
Are calculated for all i (step S107), and the maximum value Dmax among them is obtained (step S108).

  The maximum value Dmax thus obtained is compared with a preset threshold value Gth (step S109), and if it is larger than the threshold value Gth, it is determined that the image is not significant (step S110). Otherwise, it is determined that the image is significant (step S111). When the maximum value Dmax exceeds the threshold Gth in at least one of the set line segments, it is determined that the image is not significant regardless of the results of other line segments. The determination result is displayed on the display unit 118 to notify the user (operator) (step S112).

  The threshold value Gth is comprehensive based on the type of liquid injected into the well (especially its refractive index and transmittance) and the amount of liquid, the optical characteristics of the cells to be observed, the image quality level required for the apparatus, etc. For example, it may be determined experimentally so that a desired image quality can be obtained using an appropriate dummy sample.

  FIG. 7 is a diagram illustrating an example of an image used for evaluation. FIG. 7 (a) is an example of an original image taken in a state where an appropriate culture solution is injected into the well W. A dark region due to the shadow of the well wall surface is seen in the peripheral portion, and the central portion is Bright but foreign material is distributed. FIG. 7B is a diagram showing a luminance profile along a white line drawn in the horizontal direction at the central portion of FIG. The luminance value for each pixel indicated by the marker “□” varies due to the influence of these foreign substances and the like, but the influence is reduced in the profile after the smoothing process indicated by the solid line.

  Further, as can be seen from FIG. 7A, the change in luminance from the central portion to the peripheral portion is relatively gradual, and the boundary between the bright region and the dark region is not clear. Reflecting this, the gradient value obtained as the absolute value of the difference between adjacent data is also relatively small. The extremely large value at the right end of the graph seems to correspond to the end of the well region itself, not due to the influence of the shadow. Such an image is determined to be significant. On the other hand, in an image in which the boundary between the bright area and the dark area is clear as shown in FIG. 3B, the value of the gradient is increased, so that it is determined that there is no significance.

  As described above, in this embodiment, the significance of an image of a well image obtained by imaging a well into which a liquid has been injected is determined based on the magnitude of change in luminance. Specifically, the luminance profile of the pixel located on the line segment set across the well region is obtained, and when the absolute value of the gradient exceeds a predetermined threshold, it is determined that the image is not significant. To do. By doing this, it is possible to detect in advance images that may be erroneously detected such as cells due to abrupt changes in the background density due to the reflection of shadows on the well wall surface, and prevent such erroneous detection in advance. it can. Further, by using such determination for evaluation of the imaging apparatus, it can be used for improvement of imaging conditions including illumination conditions and improvement of image processing technology for captured images.

  In this embodiment, a plurality of line segments in different directions are set, and a luminance profile on each line segment is obtained. If any one of them exceeds the threshold, it is determined that the image is not significant regardless of the results of other line segments. Thereby, even when the shadow of the well wall surface appears in the image with anisotropy, the probability of overlooking the influence can be reduced.

  In addition, with respect to the luminance profile, smoothing processing and sampling processing are appropriately performed on the data in units of pixels, thereby preventing evaluation errors due to foreign matters or bubbles reflected in the image.

  As described above, in this embodiment, the microplate M corresponds to the “sample holding plate” of the present invention, and the well W corresponds to the “concave portion” of the present invention. Further, steps S101 and S102 in FIG. 5 correspond to the “image acquisition process” of the present invention, while steps S103 to S106 correspond to the “profile derivation process” of the present invention. Steps S108 to S111 correspond to the “determination step” of the present invention.

  In the above embodiment, a “luminance profile” is used as a concept corresponding to the “image density profile” of the present invention. Although there is a difference that the luminance value becomes smaller as the image density is higher, it is obvious that the luminance is an index of the image density, and in this embodiment, the “luminance profile” is substantially obtained in the present embodiment. This corresponds to obtaining the “image density profile” of the invention.

  The present invention is not limited to the above-described embodiment, and various modifications other than those described above can be made without departing from the spirit of the present invention. For example, the above embodiment is an imaging apparatus that executes the image evaluation method according to the present invention, but the application target of the present invention is not limited to this. For example, the image evaluation method of the present invention is executed by a personal computer or workstation, and the image data captured by another image capturing apparatus is captured and evaluated, thereby evaluating the image itself or the image capturing apparatus capturing the image. can do. Of course, a dedicated evaluation apparatus for executing the image evaluation method of the present invention may be configured. In other words, the “image data acquisition step” of the present invention includes not only acquiring image data by capturing an object, but also acquiring image data captured by another device via appropriate communication means. Includes.

  Further, for example, in the above embodiment, an image obtained by imaging a microplate M having a plurality of wells W having a substantially circular cross section is targeted, but the shape and number of wells (recesses) are not particularly limited. The present invention can be applied to any one.

  Moreover, although the smoothing process in the said embodiment is based on a moving average calculation, processing methods other than this are applicable. For example, as shown in FIG. 7 (a), foreign objects and the like often appear darker than the background portion in the image, and therefore smoothing may be performed by excluding pixel data that is significantly lower in brightness than adjacent data. Good. Specifically, such a smoothing process can be performed by obtaining an upper envelope of a luminance data string including variations.

  The present invention can be particularly preferably applied to a field that requires observation of a sample in which a liquid is injected into a depression, such as a well on a microplate used in the medical / biological science field. The application field is not limited to the medical / biological science field.

DESCRIPTION OF SYMBOLS 1 Imaging device 10 Control part 13 Imaging unit BR Bright part area | region DR Dark part area | region Iw Well image M Microplate (sample holding plate)
S101, S102 Image acquisition process S103-S106 Profile derivation process S108-S111 Judgment process W well (recessed part)

Claims (8)

An image data acquisition step of acquiring image data obtained by imaging the recess in a state where liquid is injected into the recess provided in the sample holding plate;
A profile derivation step of setting a straight line across the depression in the image of the depression and obtaining an image density profile in the depression along the straight line based on the image data;
Determining a gradient of the image density profile, and comparing the value with a predetermined threshold to determine the significance of the image ,
In the determination step, it is determined that the image is significant when the maximum absolute value of the gradient is smaller than the threshold value, and the image is not significant when the maximum value is larger than the threshold value. > An image evaluation method characterized by that. The image evaluation method according to claim 1, wherein in the profile deriving step, a smoothing process is performed on a data string in which pixel values of pixels on the straight line are arranged in an arrangement order to derive the image density profile. In the determination step, according to claim 1 to a value of the gradient difference between adjacent sampled data in a data string data string formed by arranging pixel values in the order of arrangement of the pixels sampled at a predetermined period in the straight line or 2. The image evaluation method according to 2. In the profile derivation step, a plurality of different straight lines are set, and the image density profile is obtained for each of the straight lines,
The determination step determines that the image is not significant if at least one of the plurality of straight lines has a maximum gradient value of the image density profile along the straight line that exceeds the threshold value. The image evaluation method according to any one of 1 to 3 . The image evaluation method according to claim 4 , wherein in the profile derivation step, the plurality of straight lines intersecting with each other in the recess are set. The image evaluation method according to claim 5 , wherein an upper surface opening of the recess is substantially circular, and each of the plurality of straight lines is set to pass through a center of the recess. In the image data acquisition step, acquiring image data corresponding to an image including a plurality of the depressions provided in the sample holding plate,
The profile deriving step and the image evaluation method according to any one of claims 1 to 6 said determining step is performed for each of the plurality of recesses. Wherein the image data acquisition process, light is irradiated from above the sample holding plate to the recess, according to any one of claims 1 to 7 for imaging by receiving light transmitted downward of the recess Image evaluation method.