Computer-implemented method for assessing the growth of germinative growth of germinated units

Abstract

A computer-implemented method for evaluating the growth of a germinant unit includes the steps of processing an image of a sample containing a germinant unit to identify at least one germinant growth present in the image that is a germinant growth of the germinant unit, determining the length of the at least one identified growth, and calculating an image average length of the at least one determined length.

Description

[Technical Field]

[0001] The present disclosure relates to a computer-implemented method for assessing the growth of germination units, and in particular, but not exclusively, to image processing for assessing the growth of germination units by determining the length of the germination units. [Background technology]

[0002] In recent years, interest has been growing in biological products (or formulations) that can replace or supplement chemical pesticides. Biological products offer the advantages of reducing risks to the environment and human health, reducing the likelihood of resistance, lowering development costs, and shortening the time required to create new products.

[0003] Fungi are one type of biological agent. The presence of fungal spores on seeds has been shown to improve seed growth by acting as natural insecticides and promoting seed development. Trichoderma is a genus of fungi that helps protect seeds on which Trichoderma spores are attached. The growth of these spores is monitored along with seed germination, and it is known that the higher the spore viability, the more likely the seeds on which the spores are attached will germinate. However, some compounds applied to seeds can affect spore growth, thereby affecting the growth of the seeds on which they are attached. One current method for assessing the effects of chemicals on biological agents such as fungal spores is the so-called viability method, or conidial germination method, which determines the percentage of viable conidia (conidia that germinate) when exposed to a specific compound. The development of a method that can evaluate the effects of compounds or chemical products on germinating units such as spores is desirable. Summary of the Invention

[0004] According to one aspect of the present invention, a computer-implemented method for assessing the growth (e.g., vigor) of germinative units is provided, comprising the steps of processing an image of a sample containing germinative units to identify (detect) at least one germinative outgrowth present in the image that is a germinative outgrowth of the germinative unit, determining the length of the identified at least one outgrowth, and calculating an image average length of the determined length. The processing step may also include identifying at least one germinated germinative unit. The vigor or growth of the germinative unit may be determined based on the average length of the sample. The average length of the germinative unit outgrowth may be correlated to the vigor of the germinative unit. For example, the longer the average length of the sample, the higher the vigor of the sample. The vigor may be determined based on the average length of samples grown under specific conditions, and the average length may be compared to a length (benchmark) value for known vigor of a particular type of sample.

[0005] The method may further comprise processing a plurality of images of the sample containing the germinated units and averaging the image mean lengths of the resulting plurality of images to generate a sample mean length.

[0006] Identifying growths in the image may include processing the image to distinguish between the image background and growths (objects, foreground) in the image. Clustering may be used to distinguish between growths and the image background. Thresholding the image may distinguish between growths and the image background. The processing may include generating a binary image based on the image. The processing may include generating a first binary image and a second binary image using different processes. The first binary image and the second binary image may be used individually or in combination to detect objects in the image.

[0007] The image may be processed to identify growths by identifying connected components in the image. Identifying growths in the image may further include performing morphological closing on the connected components. Connected components with an area less than a threshold may be determined to be at least one of ungerminated germination units and dirt (e.g., soil particles). At least one of ungerminated germination units and dirt may be ignored in calculating the image average length. The location of the germination units may be determined by detecting approximately circular regions in the image. The length of the growth in the image may be determined by determining the number of pixels associated with the growth and dividing the number of pixels of the growth by the width of the growth.

[0008] The method may further comprise determining the number of growths in the image based on the number of connected components. In calculating the image average length, identified growths that touch an edge of the image may be ignored.

[0009] The germinating unit may be a spore and the outgrowth may be a germ tube. The germinating unit may be a seed and the outgrowth may be a radicle or root.

[0010] The method may further include determining the effect of the compound to which the germinating units are exposed on the growth (vigor) of the germinating units based on the image average length. For example, germinating units may be exposed to a particular chemical, which may affect their vigor. Thus, by determining the average length of germination growth in a sample exposed to a particular chemical, the average length may be compared to a standard or reference value for the sample (e.g., including germinating units of the same type exposed to the same conditions except for exposure to the chemical), and it is possible to determine how the chemical affects the vigor of the germinating units based on the relative lengths of growth.

[0011] According to another aspect of the present invention, there is provided a data processing apparatus comprising a processor for performing the steps of a method.

[0012] According to yet another aspect of the present invention there is provided a computer program comprising instructions which, when executed by a computer, cause the computer to perform the steps of a method.

[0013] According to yet another aspect of the present invention, there is provided a computer readable (storage) medium containing instructions which, when executed by a computer, cause the computer to perform the steps of a method.

[0014] Some compounds do not directly affect the viability of the germinated units, but are detrimental to the vigor of the germinated units, or their ability to germinate under suboptimal conditions, grow at a normal rate, and produce defect-free germinated outgrowths. Vigor may be determined using the length of the germinated outgrowth; the longer the germinated outgrowth, the greater the vigor.

[0015] By measuring the length of germ outgrowth (e.g., the length of the conidial germ tube) using the methods described herein, it is possible to determine which formulation components and / or products are best used to improve the likelihood of germ outgrowth survival. Additionally, the methods described herein allow for rapid and accurate determination of germ outgrowth vigor, allowing for more accurate determination of the effect of compounds on the vigor of germinated units.

[0016] It should be noted that, in this specification, certain features will be described with respect to only one or a few aspects or embodiments of the invention to avoid unnecessary duplication of effort and repetition of text. However, it should be understood that, where technically feasible, features described in any aspect or embodiment of the invention may also be used in other aspects or embodiments of the invention. [Brief explanation of the drawings]

[0017] For a better understanding of the present invention, and to show more clearly how the same may be carried into effect, reference will now be made, by way of example, to the accompanying drawings in which:

[0018] [Figure 1] FIG. 1 is a flow diagram illustrating a method according to one embodiment. [Figure 2] FIG. 1 shows two images of a sample taken at different resolutions. [Figure 3] FIG. 10 shows three images selected for a sample according to one embodiment. [Figure 4] FIG. 1 is a flow diagram illustrating a method according to one embodiment. [Figure 5] 1 is an image showing an image generated in a step of a method according to an embodiment; [Figure 6] FIG. 1 shows an original image of a sample at 100x magnification along with images generated during steps of a method according to one embodiment. [Figure 7] FIG. 1 shows an original image of a sample at 40x magnification, along with images generated during steps of a method according to one embodiment. [Figure 8] 1 is a graph showing the average number of germinated germination units and the average length of growth from the germinated units according to one embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0019] The methods described herein generally relate to determining the germination growth length of germinative units, such as germinating units, e.g., several dispersal units of fungi or plants, such as spores or seeds. The determined length can be used to determine how a chemical (or biological) agent affects the growth of the germinative units. In one specific example, the germinative units may be fungal spores, and the germination growth length of the spores can be determined to indicate the vigor of the spores (e.g., their ability to germinate under suboptimal conditions, grow at a normal rate, and produce an intact germinative growth (germ tube)). The spores can be applied to seeds, and the vigor of the spores relates to their ability to protect the seeds from compounds or chemicals (e.g., so that the seeds germinate). By determining the germination growth length of the spores, for example, when the spores are exposed to a particular compound or chemical, the likelihood of the seeds in which the spores are exposed to the compound or chemical can be determined.

[0020] In yet a further embodiment, the germination unit may be a seed, and the length of the seed germination growth may be determined to indicate seed vigor, or the seed's ability to germinate under suboptimal conditions, grow at a normal rate, and produce an unimpaired germination growth (radicle or root). For example, if the seed is exposed to a particular compound or chemical, the length of the seed germination growth may be determined to determine the effect of the compound on seed germination. In this manner, the method described herein may also be used to analyze the seed itself to determine the effect of the compound on seed vigor. For example, the same method described herein may be applied to analyze the radicle or root germination growth of a plant from a seed. This method may also be used to determine how well the plant's radicle or root grows under different growth conditions (temperature, amount of water, etc.) by measuring the length of the radicle or root germination growth.

[0021] Therefore, the method described herein is generally used to determine the vitality of germinative units. The method described herein relates to a computer-implemented method for assessing the growth or vitality of germinative units, and includes the steps of processing an image of a sample containing germinative units to identify at least one germinative growth present in the image that is a germinative growth of the germinative unit, determining a length of the identified at least one growth, and calculating an image average length of the at least one determined length in the image. The steps of this method are shown in FIG. 1 , which shows step S101 of processing an image of a sample containing germinative units to identify at least one germinative growth present in the image that is a germinative growth of the germinative unit, determining a length of the identified at least one growth, and step S102 of calculating an image average length of the determined lengths.

[0022] The images used in this method may be images of samples containing germinating units in which germinal growth can be seen. For example, the images may be images of fungal spores with germ tubes growing from them, or images of seeds with radicle or root germination. While the images used in this example are images of germ tube growth from fungal (particularly Trichoderma) spores, these methods are equally applicable to analyzing radicle or root growth, and so the images used in these methods may be images of seeds with radicle or root germination.

[0023] For spores, especially Trichoderma, the samples used in this method were plated at 10 4 or 10 5 The dilution may be adjusted by growing spores at a dilution of 1000 kJ / ml. This dilution can reduce crossover or clumping of germination growth. The incubation period of the plates may be the same for all samples so that similar comparisons can be made. For example, a recommended incubation period is 15 hours, which is particularly suitable for Trichoderma. The dilution and incubation period may be selected appropriately for different types of germination units. Samples may be exposed to specific compounds to determine the effect of the compounds on the vitality of the germination units.

[0024] Images may be acquired using an optical microscope equipped with a high-resolution built-in camera. Images may be acquired at different resolutions, and the methods herein may be performed on these images. Note that while the images herein were taken at 40x (×40) and 100x (×100) magnifications, any magnification that allows observation of the germinating growth of the germinating units may be used. In particular, different magnifications may be used depending on the type of growth (e.g., spores require higher magnification than radicles or roots).

[0025] The images used in this method may be selected based on the number of clumps or crossovers of germination growth in the image. Figure 2 shows two example images containing clumps and / or crossovers of germination growth. For example, Figure 2 shows multiple germination growths 210 of spores 212, and also shows a clump 214 of spores and an example crossover 216. A clump is an area where multiple spores have germinated closely together, making it difficult to distinguish between the germination growths of these spores. A crossover is an area where the germination growths of two spores intersect. In either case, it is difficult to analyze the germination growths individually, for example, to determine the length of each germination growth. Therefore, analyzing an image containing many such clumps or crossovers may reduce the accuracy of the calculated average length of the germination growth. Therefore, it is beneficial to select images with few or no crossovers and / or clumps. The images may be selected by the user.

[0026] Figure 3 shows three example images containing Trichoderma spores that may be selected for use in the present method. Figures 3a and 3b are images of samples taken at 100x magnification, and Figure 3c is an image of a sample taken at 40x magnification. While these images are shown here in grayscale, images taken with an optical microscope may be in color, and it is assumed that images processed by the method described in this example will be color images. Grayscale images may also be used. Sprouting growth 310 can be seen in Figures 3a and 3c. However, no sprouting growth is visible in Figure 3b.

[0027] The method according to this embodiment will now be described. Figure 4 shows an overview of the steps of the method. Figure 5 shows an image obtained by the method according to the embodiment.

[0028] In step 1 (corresponding to S401 in FIG. 4), an image (or a group of images, e.g., an image folder) to be processed is selected (original image). In this example, an image of a sample taken at 100x magnification, as shown in FIG. 3a, is used for processing. This image is shown in FIG. 5a.

[0029] In step 2 (corresponding to S402 in FIG. 4), the background of the image is removed, returning a binary image. The returned image is shown in FIG. 5b. As can be seen from this image, pixel values ​​are replaced with either 1 or 0 to generate a black-and-white representation of the image of FIG. 5a. White pixels are generally replaced with pixels that indicate sprout growth, and black pixels are generally replaced with the background. Thresholding may be used to generate the image, where one range of pixel values ​​results in black pixels and another range of pixels results in white pixels. In step 2, a first binary image and a second binary image may be generated (not shown in FIG. 5). The first binary image and the second binary image may be generated using different image processing techniques. The first binary image and the second binary image may be used together in subsequent steps of the method to extract different information about the sprout growth in the image or to improve the relative accuracy of the information extracted from each image. The first binary image and the second binary image may be generated and / or used differently depending on the magnification of the image or the size of the growth. Alternatively, a second binary image may be generated in step 3.

[0030] In step 3 (corresponding to S403 in FIG. 4), it is determined whether or not sprouts are present in the binary image. This can be achieved by determining the ratio of white pixels to black pixels in the binary image. If the number of white pixels to black pixels exceeds a threshold, it can be determined that sprouts are not present in the image (see, for example, FIG. 6d, where no sprouts are present). If no sprouts are present in the image, the image may be ignored (e.g., by assigning a null length and null number for sprouts for images in which no sprouts are detected). The image used in this example was not discarded because some sprouts were present.

[0031] Step 4 (corresponding to S404 in FIG. 4) involves removing the remaining background (white areas of the image that are not sprouts or sprout units). As shown in FIG. 5b, the image includes a plurality of sprouts 510, generally indicated by white lines, and several roughly circular white components. These roughly circular components 518 typically include dirt (e.g., soil) or unsprouted (no sprouts attached) sprout units. These small, roughly circular components 518 that form the remaining background can be removed from the image. For example, because the remaining background typically consists of small areas of white pixels, any detected components of the image that consist of fewer than a threshold number of white pixels may be removed from the image. This method may also remove all unsprouted (no sprout units attached) sprout units in the image.

[0032] The location of germinated units from which growths have germinated can also be determined. For example, the image can be processed to determine roughly (or substantially) circular regions of pixels. Generally, these roughly circular regions of pixels correspond to germinated units from which growths have germinated. The roughly circular regions can be elliptical or have an uneven boundary that, when smoothed, results in a circular or elliptical shape. Because the circular regions can be roughly circular, they also encompass roughly circular shapes that are elliptical or more irregularly shaped (thus, this method is intended to detect circular spores or more elliptical seeds). Figure 5c shows an image in which the remaining background has been removed and circles of white pixels have been detected. As can be seen from this image, each circle of white pixels is represented by a ring 522 surrounding each circle. Figure 5c shows a magnified portion (indicated by a dotted square) containing one such ring 522. In this manner, the number of germinated units with germinated growths in the image can be determined.

[0033] Step 5 (corresponding to S405 in FIG. 4) removes any sprouting growths that touch the edges of the image. Removing edge-touching growths is beneficial because the length of the growths determined in the image is not representative of the actual length of the growths, since the entire length of the growths is likely not visible in the image. When these growths are included in the calculation of the average length of the growths, the average length may not be representative of the length of growths actually present in the sample. The image is processed to detect growths that touch the edges of the image, and these growths are then removed from the image. The result of this process is shown in FIG. 5d, and when compared to the image shown in FIG. 5c, it can be seen that the sprouting growths have been removed from this image (the area where the growths have been removed is shown by the dotted circle 520).

[0034] In step 6 (corresponding to S406 in FIG. 4), the total length of the sprouts in the image is determined. The pixels occupied by the sprout units found in step 4, S404, may be masked, and the remaining white pixels may be counted to determine the number of white pixels corresponding to the sprouts in the image. The width of the sprouts may be considered to be a constant width at a particular magnification, and the width may be predetermined (e.g., a measurement of the width of a typical sprout in the image may be obtained at a particular magnification, or the average width of the sprouts at a particular magnification may be determined). To determine the total length of the sprouts in the image, the number of pixels of each sprout may be divided by the width. This method is advantageous because the length can be determined even if the sprouts are disordered.

[0035] In step 7 (corresponding to S407 in FIG. 4), the average length of the sprouts in the image and the number of germinated units that have germinated are determined for the image. The average length of the sprouts can be determined by summing the lengths determined in step 6, S406 and dividing by the number of germinated units corresponding to the sprouts used to determine the total length in step 5, S405 (the number of germinated units remaining after growth bordering the edge has been removed). The number of germinated units is the number of germinated units determined in step 4, S404 (including all germinated units that have sprouts that border the edge of the image).

[0036] Steps 1 to 7 (S401 to S407) may be repeated for multiple images of the sample. At least two images may be used. For example, 10 images of the sample may be used in the above method. In this way, the average germination growth length and number of germinated germinated units per sample can be determined by dividing the sum of the average germination growth length per image and the sum of the number of germinated germinated units per image by the number of images. In this way, the sample average germination growth length and sample average number of germinated germinated units can be determined. The vitality of the germinated units can be determined by using these sample average values.

[0037] FIG. 6 shows an image taken at 100x magnification selected in step 1 of the method. The original images are shown in FIGS. 6a and 6b. FIGS. 6c and 6d show first binary images of the images of FIGS. 6a and 6b, respectively, generated in step 2 of the method. FIGS. 6e and 6f show second binary images of the images of FIGS. 6a and 6b, respectively, generated in step 2 of the method. As can be seen from FIG. 6, images 6a and 6b are processed differently to generate two different binary images, the first binary image and the second binary image. For example, in this example, second binary images 6e and 6f are generated by modifying the contrast of images 6a and 6b and applying a standard deviation filter before binarization. These methods result in a second binary image with exaggerated growth, making the image clearer and more useful for subsequent processing. In step 3, it may be determined whether or not sprout growth is present in the binary image. In this example, based on binary images 6c and 6e, it is determined that sprouts are present in the image corresponding to Figure 6a, but based on binary images 6d and 6f, it is determined that sprouts are not present in the image corresponding to Figure 6b. Therefore, the image corresponding to Figure 6b is ignored for the remainder of the method. In step 4, the first binary image 6c and the second binary image 6e are used to remove residual background from the first binary image 6c. For example, the residual background is detected in the second binary image 6e and then used to mask the corresponding pixels in Figure 6c. In step 4, the locations of spores from which growth had germinated were also determined, so the number of spores with sprouts can be determined. Figure 6g shows the resulting image after removing the residual background and the locations of the spores in the image. In step 5, all sprouts that touch the edges of the image are removed. The result of this removal is shown in Figure 6h. Steps 6 and 7 may then be performed to determine the average length of the sprouts in the image of Figure 6h.

[0038] In this example, the number of germinated spores detected was 9, and the average length of the germ tube was 918.48 pixels, corresponding to 192.15 μm.

[0039] The method for generating the image shown in Figure 6 is described in more detail below. This method is particularly useful at 100x magnification for determining the average length of Trichoderma spores.

[0040] In step 1, raw images of the sample are selected for analysis. Images may be selected based on the number of sprouts in the image. Images may also be selected based on the distribution of sprouts in the image. For example, images with fewer clumps of growth compared to other images may be selected. In this example, two images corresponding to the images in Figures 3a and 3b are selected.

[0041] In step 2, a first binary image is generated for each original image selected in step 1. The first binary image is created by converting the sRGB values ​​of each pixel in the image to LAB color space values. The resulting data may then be converted to single precision and image segmentation using K-means clustering may be performed. The pixel image may be divided into two clusters of pixels (each cluster corresponding to pixels defining a sprout growth) using clustering. The clustering may be repeated three times, for example. Clustering can separate the background from the foreground (or objects in the image) even when image quality is poor (e.g., due to darkening or lightening of areas in the image that should have the same shade). The segmented images may then be binarized using a binarization function to generate the first binary image. For example, the first binary image may be created by replacing all pixel values ​​above a globally determined threshold with 0 (black) and all pixel values ​​below the threshold with 1 (white). In this way, pixels thresholded to correspond to sprout growth may be white pixels, and pixels thresholded not to correspond to sprout growth may be black pixels. The threshold may be determined using a method such as Otsu's method, in which the threshold is selected to minimize the intra-class variance of the thresholded black and white pixels.

[0042] In step 3, a pre-computation may be performed to determine whether or not sprouts are present in the image. Specifically, a histogram of pixel values ​​of the first binary image created in step 2 may be generated. The number of black pixels and white pixels may be determined. Based on the determined number of black pixels and the number of white pixels, the ratio of white pixels to black pixels may be calculated.

[0043] If the calculated ratio is greater than a first threshold, e.g., greater than 10, a complement of the first binary image may be generated in which the pixel values ​​are inverted (the image is black and white). The number of black and white pixels in the complemented binary image may then be determined, and a new ratio of white to black pixels may be calculated. The ratio of black to white pixels depends on the contrast of the image. Also, if the image is too dark (the number of black pixels is greater than the number of white pixels), the binarization may assign 1 to the background and 0 to the object in the image, rather than assigning 1 to the object and 0 to the background. In this case, the complement of the first binary image may invert 0s and 1s so that the object has a value of 1 and the background has a value of 0.

[0044] If the calculated ratio, or a newly calculated ratio where the calculated ratio is greater than the first threshold, is greater than a second threshold, it is determined that there are no germinated units in the image. For example, if the number of white pixels is greater than the number of black pixels, it may be determined that there are no germinated units in the image. In this case, the image is assigned a value of 0 for the average length and a value of 0 for the number of germinated units. Alternatively, the original image may be processed to detect germinated units that have not germinated. Specifically, the sRGB values ​​of each pixel in the original image may be converted to a grayscale image and then processed. In a grayscale image, germinated units that have not germinated can be more easily detected.

[0045] If the calculated ratio, or a newly calculated ratio where the calculated ratio is greater than the first threshold, is less than a second threshold, e.g., less than 0.1, a second binary image may be created from the original image. The second binary image may be created by converting the sRGB values ​​of pixels in the original image to a grayscale image. Image intensity values ​​are then adjusted, e.g., by saturating the bottom 1% and top 1% of all pixel values ​​to increase contrast. A standard deviation filter is applied so that the value of each output pixel is the standard deviation of a neighborhood of the corresponding input pixel. The neighborhood may, for example, be an n-by-n matrix of ones, where n is 25. Symmetric padding may be used for pixels on the border of the image. In symmetric padding, the values ​​of the padding pixels are the specular reflections of the border pixels in the image. The image may then be processed to generate its complement. The complement of the image may then be binarized. Binarization may include converting the image to grayscale and using adaptive thresholding with a sensitivity factor, e.g., a sensitivity factor of 0.7. The resulting image may then be processed to produce its complement, which may be a second binary image.

[0046] If the calculated ratio, or a new calculated ratio where the calculated ratio becomes greater than the first threshold, is less than the second threshold, the method proceeds to step 4.

[0047] In step 4, the second binary image is processed to find connected components in the second binary image, which can form individual objects (e.g., sprouts). For example, pixels adjacent to pixels with the same value are likely to be the same object. Pixels can be considered connected if their edges touch each other (called 4-connectivity (or connectivity)). In this case, two adjacent pixels are part of the same object if they are both horizontally or vertically aligned and connected along the horizontal or vertical direction. Alternatively, pixels can be considered connected if their edges or corners touch each other. In this case, two adjacent pixels are part of the same object if they are both horizontally, vertically, or diagonally aligned and connected along the horizontal, vertical, or diagonal direction (also called 8-connectivity). In this example, 4-connectivity is used. 4-connectivity is also useful in preventing objects that are close to each other from being determined to be the same object.

[0048] The number of objects detected in the second binary image may be determined. Properties such as shape measurements and pixel value measurements may also be calculated for the objects in the second binary image. Information about the shape and size of the detected objects may be used to remove all ungerminated germination units and dirt (as used herein, dirt refers to any object present in the image that is not a germination unit or sprout growth; for example, soil particles may be present in the image as dirt) from the image. Specifically, objects detected in the second binary image with an area equal to or less than a threshold may be removed from the first binary image. For example, if the area of ​​an object is 2000 pixels or less, the object may be considered an ungerminated germination unit or a dirt in the sample. These objects may be removed from the first binary image because they are not important. Also, ungerminated germination units and dirt may be removed from the first binary image by finding circular objects or substantially circular objects with an eccentricity equal to or less than a threshold. This allows small substantially circular objects to be removed, while small germination tubes that appear circular but have an eccentricity greater than a threshold cannot be removed from the binary image.

[0049] The first binary image may then be processed to find connected components in the same manner as the second binary image. The first binary image is processed after the objects (ungerminated sprout units and dirt) detected in the second binary image are removed from the first binary image. In this manner, the first binary image may be processed to determine the number of objects detected in the first binary image, and shape measurements and pixel value measurements of the objects in the first binary image may be calculated as described above. The first binary image may also be processed to determine the presence or absence of ungerminated sprout units or dirt remaining in the first binary image. For example, if an object has an area of ​​300 pixels or less, the object may be considered to be an ungerminated sprout unit or dirt in the sample. These objects may be removed from the first binary image because they are not important.

[0050] The location of a germinating unit from which growth has germinated may be detected in the first binary image. The germinating unit may be located by detecting circles, for example, by using a function that detects circles using a Hough transform. The location of the circle's center may also be determined. The germinating unit may be detected by detecting circles with a radius in a range of, for example, 11 to 20 pixels and using adaptive thresholding with a sensitivity factor, for example, a sensitivity factor of 0.9 (to allow for some eccentricity in the circle shape). The detected circle locations may be stored in an array with a number of columns and rows corresponding to the pixels in the first binary image. It may then be determined whether all pixels within the defined circle have the same pixel value (e.g., whether all pixels are white). If all pixels within the circle have the same value (e.g., a value of 1), the circle is determined to correspond to a germinating unit. A mask of pixels corresponding to the germinating unit may be created.

[0051] The detection of germination units may be repeated by performing similar steps while varying the range of possible circle radii. For example, a subsequent step may detect circles with radii between 14 and 25 pixels. This step may identify more circles with improved correspondence to existing germination units in the image. The circles detected in this step are compared with the circles determined in the previous step, and overlapping circles are discarded.

[0052] Next, in step 5, all growths that touch the edges of the image are removed from the image. Removing boundary-connected growths is advantageous because it is likely that at least part of the growth will not be present in the image, and therefore the lengths determined from the image will not be representative of the actual lengths of the growths in the sample. This avoids determining growth lengths that are not representative of the actual growth lengths, which could result in a misalignment of the average growth lengths in the image.

[0053] Structures bordering the image edges may be removed by suppressing structures that are brighter than their surroundings and connected to the image boundary. In this method, a clean image may be generated from a second binary image in which structures connected to the image boundary have been removed using 8-connectivity. The clean image may then be subtracted from the second binary image to create a mask by filling image regions and holes in the second binary image (a hole is a set of background pixels that cannot be reached by filling the background from the image edge). All growths bordering the edge are filled in because they have the same pixel value as the boundary.

[0054] A mask may be applied to the first binary image, and a clean image may be generated from the first binary image in which structures connected to the image boundary have been removed, as described above. The number of white pixels in the cleaned first binary image may be determined and compared to a threshold, and if the number of white pixels is less than the threshold, e.g., 0.3 times the number of white pixels in the first binary image before the boundary was cleared, the first binary image is reverted to the first binary image before step 5. In this way, if too many pixels have been removed, all pixels are retained. Thus, it is determined whether too many pixels have been removed, and if so, all pixels are retained.

[0055] If the radius ranges of the detected pixels are different, the remaining germination units may be detected by repeating the above-described method. Also, pixels corresponding to the germination units may be masked in the first binary image. Duplicately detected germination units may be discarded.

[0056] In step 6, the number of detected germinated units may be used to determine the number of germinated units detected in step 4. The length of the germinated growth in the image may also be determined. For example, the number of pixels constituting the germinated growth in the image may be determined by masking the pixels occupied by the germinated units and counting the number of remaining white pixels. In this manner, the number of pixels constituting the germinated growth area can be determined. The width of the germinated growth may be assumed to be constant at a particular magnification, and the width may be determined by measuring the number of pixels across the entire width of the growth at a particular magnification. In this example, the average growth width is 15 pixels.

[0057] The total length of the growth in the image can then be calculated by dividing the number of pixels in the growth by the growth width. This method makes it possible to calculate the total length of the growth in the image even if the growth lengths are irregular.

[0058] In step 7, the average length of the sprouts in the image is determined. The average length of the sprouts can be determined by summing the lengths determined in step 6 and dividing by the number of germinated germ units determined in step 5 (the number of germinated units remaining after edge growth is removed). The number of germinated germ units is the number of germinated germ units determined in step 4.

[0059] FIG. 7 shows an image taken at 40x magnification selected in step 1 of the method. The original image is shown in FIG. 7a. FIG. 7b shows a first binary image of the image in FIG. 7a generated in step 2 of the method. FIG. 7c shows a second binary image of the image in FIG. 7a generated in step 2 of the method. As can be seen from FIG. 7, the image in FIG. 7a is processed differently to generate two different binary images, the first binary image and the second binary image, as described above. In step 3, it may be determined whether germinated growth is present in the binary image. In this example, it is determined that germinated growth is present in the image corresponding to FIG. 7a based on the binary images shown in FIGS. 7b and 7c. In step 4, the first binary image 7b and the second binary image 7c are used to remove any remaining background from the first binary image 7b. In step 4, the locations of spores with germinated growth are also determined, allowing the number of spores with germinated growth to be determined. Figure 7d shows the resulting image after removing the remaining background and the location of the spores in the image. In step 5, we remove all spore growths that touch the edge of the image. The result of this removal is shown in Figure 7e. Steps 6 and 7 are then performed to determine the average length of the spores in the image in Figure 7e.

[0060] In this example, the number of germinated spores detected was 46, and the average length of the germ tube was 279.6 pixels, corresponding to 137.37 μm.

[0061] The sample average germination length and sample average number of germinated units may then be used to determine the vigor (or relative vigor) of the germinated units (which correlates with germination length). The effect of a compound on spore growth can be determined, for example, by comparing the sample average lengths of various samples exposed to different compounds, or by comparing with standard values ​​for the length of germinated units exposed under standard conditions.

[0062] Below we will explain in more detail how the image shown in Figure 7 was generated. This method is particularly advantageous at 40x magnification for determining the average length of Trichoderma spores.

[0063] In step 1, an original image (or images) of the sample is selected for analysis. Images may be selected based on the number of sprouts in the image. Images may also be selected based on the distribution of sprouts in the image. For example, images with fewer clumps of growth compared to other images may be selected. In this example, the image corresponding to the image in Figure 3c is selected.

[0064] In step 2, a first binary image is created for the source image selected in step 1. Contrary to the method described above for the 100x magnification, in this embodiment, the first binary image may be created by converting the sRGB values ​​of the pixels in the image to grayscale. Morphological diclosing can be performed on the grayscale image to generate a morphological diclosing image. A morphological diclosing operation involves dilation followed by erosion, using the same structuring element for both operations. Morphological diclosing can result in an image with filled gaps. In this embodiment, a disk-shaped structuring element with a radius of 25 pixels may be used. This method is particularly advantageous when the image growths are relatively small, especially when a lack of resolution causes one growth to appear as two separate growths.

[0065] A grayscale image may be obtained from the morphologically closed image. The resulting image may then be processed to generate its complement. To complement a grayscale image, each pixel value is subtracted from the maximum pixel value supported by the class (1.0 for double-precision images). The complemented image is then binarized, and the complement of the binarized image is the first binary image.

[0066] A second binary image may be created for the original image. The second binary image may be created by converting the sRGB values ​​of pixels in the original image to a grayscale image. Image intensity values ​​are then adjusted, for example, by saturating the bottom 1% and top 1% of all pixel values ​​to increase contrast. A standard deviation filter is applied so that the value of each output pixel is the standard deviation of a neighborhood of the corresponding input pixel. The neighborhood may be, for example, an n-by-n matrix of ones, where n is 9. Symmetric padding may be used for pixels on the border of the image. In symmetric padding, the values ​​of the padding pixels are the specular reflections of the border pixels in the image. The image may then be processed to generate its complement. The complement of the image may then be binarized. Binarization may include converting the image to grayscale and using adaptive thresholding with a sensitivity factor, for example, a sensitivity factor of 0.6. The resulting image may be the second binary image.

[0067] In step 3, a pre-computation may be performed to determine whether or not sprouts are present in the image. Specifically, a histogram of pixel values ​​of the second binary image created in step 2 may be generated. The number of black pixels and white pixels may be determined. Based on the determined number of black pixels and the number of white pixels, the ratio of white pixels to black pixels may be calculated.

[0068] If the calculated ratio is greater than a first threshold, e.g., greater than 50, a second binary image may be interpolated. Then, the number of black and white pixels in the interpolated binary image may be determined, and a new ratio of white to black pixels may be calculated. The ratio of black to white pixels depends on the contrast of the image. Also, if the image is too dark (the number of black pixels is greater than the number of white pixels), the binarization may assign 1 to the background and 0 to the object in the image, rather than assigning 1 to the object in the image and 0 to the background. In this case, the interpolation of the first binary image may be performed by inverting the 0s and 1s.

[0069] If the calculated ratio, or a newly calculated ratio where the calculated ratio is greater than the first threshold, is greater than a second threshold, it may be determined that there are no germinated units in the image. For example, if the number of white pixels is greater than the number of black pixels, it may be determined that there are no germinated units in the image. In this case, the image is assigned a value of 0 for the average length and a value of 0 for the number of germinated units. The raw image may be processed to detect germinated units that have not germinated. In a grayscale image, germinated units that have not germinated may be more easily detected.

[0070] If the calculated ratio, or a newly calculated ratio where the calculated ratio is greater than the first threshold, is less than the second threshold, e.g., less than 0.5, then perform step 4.

[0071] In step 4, the second binary image is processed to find connected components in the second binary image, which can then form individual objects as described above. In this example, 4-connectivity is used.

[0072] The number of objects detected in the second binary image may be determined. Object characteristics, such as shape measurements and pixel value measurements, may also be determined for the objects in the second binary image. Information about the shape and size of the detected objects may be used to remove all germination units and stains from the image. Specifically, objects detected in the second binary image with an area equal to or less than a threshold may be removed from the first binary image. For example, if an object's area is equal to or less than 1000 pixels, the object may be considered to be an ungerminated germination unit in the sample or a stain. These objects may be removed from the first binary image because they are not significant. Ungerminated germination units and stains may also be removed from the first binary image by finding circular objects or substantially circular objects with an eccentricity equal to or less than a threshold. This allows for the removal of small, substantially circular objects, while small germination tubes that appear circular but have an eccentricity greater than a threshold cannot be removed from the binary image.

[0073] The first binary image may then be processed to find connected components in the same manner as the second binary image. The first binary image is processed after the objects (ungerminated sprout units and dirt) detected in the second binary image have been removed from the first binary image. In this manner, the first binary image may be processed to determine the number of objects detected in the first binary image, and object characteristics, such as shape measurements and pixel value measurements, of the objects in the first binary image may be calculated as described above. The first binary image may also be processed to determine the presence or absence of ungerminated sprout units or dirt remaining in the first binary image. For example, if an object has an area of ​​400 pixels or less, the object may be considered to be an ungerminated sprout unit in the sample or a dirt. These objects may be removed from the first binary image because they are not significant. Ungerminated sprout units and dirt may also be removed from the first binary image by finding circular objects or substantially circular objects with an eccentricity below a threshold. This allows for the removal of small substantially circular objects, while not allowing small germ tubes that appear circular but have an eccentricity greater than a threshold to be removed from the binary image.

[0074] The location of the germinated unit from which the growth has sprouted may be detected from the first binary image by using adaptive thresholding with a sensitivity factor of, for example, 0.93 along with an edge threshold of 0.89 and an instruction to find all bright circles (circles with pixel values ​​close to white in the image within a radius range) and detecting circles within a radius range such as 6-12 pixels as described above.

[0075] Next, any growths that border the edges of the image are removed from the image, for example, structures in the image that are brighter than their surroundings (e.g., growths) and that are connected to the image border may be removed from the image border.

[0076] Structures bordering the image edges may be removed by suppressing structures that are brighter than their surroundings and connected to the image boundary. In this method, a clean image may be generated from a second binary image in which structures connected to the image boundary have been removed using 8-connectivity. The clean image may then be subtracted from the second binary image to create a mask by filling image regions and holes in the second binary image. All growths bordering the edge have the same pixel value as the boundary, so all growths bordering the boundary are filled.

[0077] A mask may be applied to the first binary image, and a clean image may be generated from the first binary image in which structures connected to the image boundary have been removed, as described above. The number of white pixels in the cleaned first binary image may be determined and compared to a threshold, and if the number of white pixels is less than the threshold, for example, 0.3 times the number of white pixels in the first binary image before the boundary was cleared, the first binary image is reverted to the first binary image before step 5. In this way, if too many pixels have been removed, all pixels are kept.

[0078] If the radius ranges of the detected pixels are different, the method described above may be repeated to detect the remaining germination units, the pixels corresponding to the germination units may be masked, or duplicated germination units may be discarded.

[0079] In step 6, the number of detected germinated units may be determined based on the number of germinated units detected in the previous step. The number of white pixels in the first binary image may be used to determine the total number of pixels that make up the germinated growth in the image. The germinated growth is assumed to have a standard width, and the width can be determined by measuring the number of pixels across the width of the growth. In this example, the average growth width is 5 pixels. The total length of the growth in the image can then be determined by dividing the total number of pixels that make up the growth by the growth width.

[0080] In step 7, the average length of the sprouts in the image is determined. The average length of the sprouts can be determined by summing the lengths determined in step 6 and dividing by the number of germinated germ units determined in step 5 (the number of germinated units remaining after edge growth is removed). The number of germinated germ units is the number of germinated germ units determined in step 4.

[0081] Figure 8 is a graph showing the results of the above method. Specifically, Figure 8 shows the average number of germinated spores detected and the average length of the germinated spores. Data was obtained from three samples containing Trichoderma spores T1, T2, and T3. For each sample, three images were taken under a microscope at 100x magnification (×100). Using the above method, the average number of germinated spores detected and the average length of the germ tubes were determined. The graph shows these results along with the standard deviation of the results, shown as a line through the top of each bar. Note that the average number of germinated spores is calculated based on integers (1, 2, 3, etc.), so a standard deviation of one or two units is not significant. The average length is a continuous measurement, meaning that there are an infinite number of values ​​per unit. As can be seen from Figure 8, the above method can consistently determine the average number of germinated spores and the average length of germinated outgrowths per sample.

[0082] In any of the above aspects, various features may be implemented as hardware or as software modules running on one or more processors. Features of one aspect may also be applicable to any of the other aspects.

[0083] The present invention also provides a computer program or computer program product for carrying out any of the methods described herein, as well as a computer readable medium having recorded thereon a program for carrying out any of the methods described herein. The computer program of the present invention may be stored on a computer readable medium, or may be in the form of a downloadable data signal, for example provided from an internet site, or in other forms.

[0084] A computing device, such as a data storage server, may embody the present invention or perform the methods of embodiments of the present invention. The computing device may include a processor and memory. The computing device may also include a network interface for communicating with other computing devices, such as other computing devices of embodiments of the present invention.

[0085] For example, an embodiment may consist of a network of such computing devices, which may also include input mechanisms such as a keyboard and mouse, and a display unit such as a monitor, and each component may be connectable via a bus.

[0086] Memory may include computer-readable media, which refers to a single medium or multiple media (e.g., centralized or distributed databases and / or associated caches and servers) configured to carry computer-executable instructions or store data structures. Computer-executable instructions may include, for example, instructions and data that can be accessed by a general-purpose computer, a special-purpose computer, or a special-purpose processing device (e.g., one or more processors) to cause it to perform one or more functions or operations. As such, a "computer-readable medium" may be any medium capable of storing, encoding, or carrying a set of machine-executable instructions, thereby causing the machine to perform any one or more of the methods disclosed herein. Thus, "computer-readable medium" includes, but is not limited to, solid-state memory, optical media, magnetic media, and the like. Such computer-readable media include, but are not limited to, non-transitory computer-readable storage media including, for example, random access memory (RAM), read-only memory (ROM), electrically programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disk storage, magnetic disk storage or other magnetic disk devices, and flash memory devices (e.g., solid-state memory devices).

[0087] A processor may control a computing device, perform processing operations, and implement the methods described herein, for example, by executing code stored in memory. The memory may store data that is read from and written to by the processor. Here, a processor may include one or more general-purpose processing devices, such as a microprocessor, a central processing unit, or the like. A processor may include a complex instruction set computer (CISC) microprocessor, a reduced instruction set computer (RISC) microprocessor, a very long instruction word (VLIW) microprocessor, or a processor implementing other instruction sets or combinations of instruction sets. A processor may also include one or more special-purpose processing devices, such as an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), a digital signal processor (DSP), a network processor, or the like. In one or more embodiments, a processor executes instructions to perform the operations and steps described herein.

[0088] The display unit may also provide for the display of representations of data stored by the computing device, as well as displaying cursors, dialog boxes, and screens that allow interaction between a user and programs and data stored on the computing device. The input mechanism may allow a user to input data and instructions into the computing device.

[0089] Embodiments of the present disclosure, as well as various features and advantageous details thereof, will be more fully described with reference to non-limiting examples described and / or illustrated in the drawings and detailed in the specification. It should be noted that the features depicted in the drawings are not necessarily drawn to scale, and alternative embodiments may be apparent to those skilled in the art even if not explicitly described herein. Furthermore, descriptions of well-known components and processing techniques may be omitted so as not to unduly deviate from the embodiments of the present disclosure. The examples used herein are intended merely to facilitate an understanding of how embodiments of the present disclosure may be implemented and to enable those skilled in the art to implement the same. Therefore, the examples herein should not be construed as limiting the scope of the embodiments of the present disclosure, which is defined solely by the appended claims and applicable law.

[0090] It is understood that embodiments of the present disclosure are not limited to the particular methodologies, protocols, devices, apparatus, materials, applications, etc. described herein, as these may vary. It is also understood that the terminology used herein is for the purpose of describing particular embodiments only, and is not intended to limit the scope of the embodiments as recited in the claims. It is noted that as used herein and in the appended claims, the singular forms "a," "an," and "the" include plural references unless the context clearly dictates otherwise.

[0091] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the embodiments of this disclosure belong. Preferred methods, devices, and materials are described, although any methods and materials similar or equivalent to those described herein can be used in the practice or testing of the embodiments.

[0092] Although only a few exemplary embodiments have been described in detail above, those skilled in the art will readily appreciate that many variations are possible in the exemplary embodiments without significantly departing from the novel teachings and advantages of the embodiments of the present disclosure. The above-described embodiments of the present invention may be advantageously used independently of other embodiments, but may also be advantageously used in possible combination with one or more other embodiments.

[0093] All such modifications are hereby intended to be included within the scope of the embodiments of the present disclosure, as defined in the following claims. In the claims, means-plus-function clauses are intended to cover the structures described herein as performing the recited function, and not only structural equivalents but also equivalent structures as well.

[0094] Furthermore, any reference signs placed in parentheses in one or more claims shall not be construed as limiting the claim. The use of words such as "comprise", "include", and the like does not exclude the presence of elements or steps other than those listed in any claim or the specification as a whole. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements, and vice versa. One or more of the embodiments may be implemented by hardware comprising several distinct elements. In a device or apparatus claim enumerating several means, several of these means may be embodied by one and the same item of hardware. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that these measures cannot be used to advantage in combination.

Claims

1. 1. A computer-implemented method for assessing vigor of germinating units, comprising: processing an image of a sample including a germinant unit to identify at least one germinant present in the image, the germinant being an outgrowth of the germinant unit, and determining a length of the identified at least one germinant; Calculating an image average length of the determined length; and determining an effect of a compound to which the germinated unit is exposed on the vigor of the germinated unit based on the image average length.

2. 2. The computer-implemented method of claim 1, further comprising processing a plurality of images of a sample containing germinating units and averaging the image average lengths of the resulting plurality of images to generate a sample average length.

3. 3. The computer-implemented method of claim 1, wherein identifying the at least one sprouted growth in the image comprises processing the image to distinguish between a background of the image and the at least one sprouted growth in the image.

4. The computer-implemented method of claim 3 , wherein clustering is used to distinguish the at least one sprout from the background of the image.

5. 5. A computer-implemented method according to claim 3 or 4, wherein the image is thresholded to distinguish the at least one sprout from the background of the image.

6. The computer-implemented method of any one of claims 1 to 5, wherein processing comprises generating a binary image based on the image.

7. A computer-implemented method according to any preceding claim, further comprising processing the image to identify the at least one sprout by identifying connected components in the image.

8. The computer-implemented method of claim 7 , wherein identifying the at least one sprouting growth in the image further comprises performing morphological closing on the connected components.

9. The computer-implemented method of claim 7 or 8, wherein the connected components having an area less than a threshold are determined to be at least one of ungerminated germination units and stains.

10. The computer-implemented method of claim 9 , wherein at least one of the non-germinated germinated units and smudges is ignored in calculating the image average length.

11. A computer-implemented method as described in any one of claims 7 to 10, further comprising a step of determining the number of sprouts in the image based on the number of connected components.

12. The computer-implemented method of any one of claims 1 to 11, wherein the location of a germinating unit is determined by detecting a substantially circular area in the image.

13. A computer-implemented method described in any one of claims 1 to 12, wherein the length of the at least one sprout growth in the image is determined by determining the number of pixels associated with the at least one sprout growth and dividing the number of pixels of the at least one sprout growth by the width of the sprout growth.

14. The computer-implemented method of any preceding claim, wherein the calculation of the image average length disregards identified sprouts that touch an edge of the image.

15. The computer-implemented method of any one of claims 1 to 14, wherein the germinant unit is a spore and the at least one germinal outgrowth is a germ tube.

16. The computer-implemented method of any one of claims 1 to 14, wherein the germinating unit is a seed and the at least one germinating outgrowth is a radicle or a root.

17. A data processing apparatus comprising a processor for performing the steps of the method according to any one of claims 1 to 16.

18. A computer program comprising instructions which, when executed by a computer, cause the computer to carry out the steps of the method according to any one of claims 1 to 16.

19. A computer readable medium comprising instructions which, when executed by a computer, cause the computer to perform the steps of the method of any one of claims 1 to 16.

Images