Normalization processing method and device for image brightness and electronic equipment

Abstract

The invention discloses an image brightness normalization processing method and device and electronic equipment. The method comprises the steps of obtaining a target image; according to the target image, obtaining a target parameter corresponding to the to-be-detected sample on a current image channel of the target image; the target parameter is used for representing whether the to-be-detected sample meets a base balance condition on the current image channel or not; under the condition that the target parameter represents that the to-be-detected sample meets the base balance condition on the current image channel, obtaining a normalization parameter of the target image on the current image channel at least according to a pixel point brightness value of the target image on the current image channel; under the condition that the target parameter represents that the to-be-detected sample does not meet the base balance condition on the current image channel, obtaining a normalization parameter of the target image on the current image channel according to the pixel point brightness values of the target image on all image channels; and according to the normalization parameter, carrying out normalization processing on the brightness value of the pixel point on the current image channel.

Description

Technical Field

[0001] The present application relates to the field of image processing technology, and in particular to a method, device and electronic device for normalizing image brightness. Background Art

[0002] Sequencing technology refers to: using high-resolution microscopic imaging to obtain the signal of fluorescent molecules, and obtaining the base sequence by decoding the image signal. Different sequencing platforms will have different optical systems and different fluorescent dyes to obtain different point light source signal images. In the sequencing process, there are multiple pictures of the same scene. First, the pictures of the same scene are aligned through positioning and registration algorithms, and then the point signals are extracted through algorithms. Subsequent brightness information analysis and processing is performed to obtain the base sequence. With the development of sequencing technology, sequencer products are equipped with real-time processing and analysis software for sequencing data. First, different base signals are matched, and because different lasers excite different fluorescent lights, the brightness is not uniform. Some channels have high excitation efficiency and the overall brightness is high, while some channels have low excitation efficiency and the overall brightness is low. Therefore, it is unfair to use the brightness of the original image for comparison. Therefore, normalization is required to compare the brightness of different channels on the same scale.

[0003] The existing image registration algorithm normalizes the intensity extracted from the images of each channel in a scene, sorts them from small to large according to the brightness value, selects the 30th percentile brightness value as the background and the 99th percentile brightness value as the signal, and linearly normalizes the brightness value to 0-1, eliminating the different luminous brightness of each channel caused by different laser excitation efficiency, and compares the brightness on the same scale to determine the base category.

[0004] However, this algorithm will independently normalize each channel of the base signal image based on fluorescent molecules. When the base content of the current channel is low, the luminous points are sparse and the overall brightness is low. In extreme cases, there is only background. Then, according to the previous method of normalizing each channel separately, the background will be forcibly normalized to the signal, resulting in an artificially high brightness value after normalization, which will cause serious errors in the base recognition results. Therefore, a technical solution to improve the accuracy of image brightness normalization processing is urgently needed. Summary of the invention

[0005] In view of this, the present application provides a method, device and electronic device for normalizing image brightness, so as to improve the accuracy of image brightness normalization processing, as follows:

[0006] A method for normalizing image brightness, the method comprising:

[0007] Obtaining a target image, wherein the target image is an image obtained by collecting an image of a sample to be tested under laser irradiation, wherein the sample to be tested is stained with a fluorescent dye;

[0008] According to the target image, obtaining a target parameter corresponding to the sample to be tested on the current image channel of the target image; the target parameter is used to indicate whether the sample to be tested meets a base balance condition on the current image channel;

[0009] When the target parameter represents that the sample to be tested satisfies the base balance condition on the current image channel, obtaining a normalized parameter of the target image on the current image channel at least according to the brightness value of the pixel of the target image on the current image channel;

[0010] When the target parameter indicates that the sample to be tested does not satisfy the base balance condition on the current image channel, obtaining a normalized parameter of the target image on the current image channel according to the brightness values ​​of the pixels of the target image on all image channels;

[0011] According to the normalization parameter, the brightness values ​​of the pixels on the current image channel are normalized.

[0012] The above method preferably obtains the normalization parameters of the target image on the current image channel according to the brightness values ​​of the pixels of the target image on all image channels, including:

[0013] Sort the pixels of the target image on each image channel according to the brightness value respectively;

[0014] Adding the brightness values ​​of the pixels in the first quantile on all the image channels to obtain a first value;

[0015] Adding the brightness values ​​of the pixels in the second quantile on all the image channels to obtain a second value;

[0016] The first value and the second value constitute a normalization parameter of the target image on the current image channel.

[0017] The above method preferably obtains the normalization parameter of the target image on the current image channel according to the brightness value of the pixel of the target image on the current image channel, including:

[0018] Sort the pixels of the target image on each image channel according to the brightness value respectively;

[0019] Add the brightness values ​​of the pixels in the first quantile on the current image channel to obtain a third value;

[0020] Adding the brightness values ​​of the pixels in the second quantile on all the image channels to obtain a fourth value;

[0021] The third value and the fourth value constitute a normalization parameter of the target image on the current image channel.

[0022] In the above method, preferably, before sorting the pixels of the target image on each image channel according to the brightness value, the method further comprises:

[0023] According to the corresponding scaling dimension value on each image channel, the brightness value of the pixel point of the target image on each image channel is scaled respectively.

[0024] In the above method, preferably, the scaling dimension value corresponding to the target image on any target image channel is obtained by:

[0025] Obtaining multiple frames of sample images, where the sample images are images corresponding to the base sequencing device to which the target image belongs in multiple sequencing cycles;

[0026] Sort the pixel points of the sample image on each image channel according to the brightness value respectively;

[0027] Adding the brightness values ​​of the pixels on the target image channel that are sorted between the first quantile and the second quantile to obtain a fifth value;

[0028] Adding the brightness values ​​of the pixels on all the image channels that are sorted between the first quantile and the second quantile to obtain a sixth value;

[0029] For each frame of the sample image, respectively, obtaining a ratio between the fifth value and the sixth value;

[0030] The scale values ​​are averaged according to the number of the sample images to obtain the corresponding scaling dimension value on the target image channel.

[0031] In the above method, preferably, the multiple ratio values ​​corresponding to the multiple frames of sample images satisfy a ratio stability condition.

[0032] The above method preferably obtains the target parameters corresponding to the sample to be tested on the current image channel of the target image according to the target image, including:

[0033] Sort the pixels of the target image on each image channel according to the brightness value respectively;

[0034] Adding the brightness values ​​of the pixels of the target image that are sorted in the first quantile on the current image channel to obtain a seventh value;

[0035] Adding the brightness values ​​of the pixels in the first quantile on all the image channels to obtain an eighth value;

[0036] At least the seventh value and the eighth value are used to obtain the target parameter corresponding to the sample to be tested on the current image channel.

[0037] In the above method, preferably, at least the seventh value and the eighth value are used to obtain the target parameter corresponding to the sample to be tested on the current image channel, including:

[0038] When the ratio of the seventh value to the eighth value is greater than or equal to the balance threshold, obtaining a target parameter characterizing that the sample to be tested satisfies a base balance condition on the current image channel;

[0039] When the ratio of the seventh value to the eighth value is smaller than the balance threshold, a target parameter is obtained that indicates that the sample to be tested does not satisfy a base balance condition on the current image channel.

[0040] A device for normalizing image brightness, the device comprising:

[0041] An image acquisition unit, used to acquire a target image, wherein the target image is an image acquired by acquiring an image of a sample to be tested under laser irradiation, wherein the sample to be tested is dyed with a fluorescent dye;

[0042] A parameter obtaining unit, used for obtaining, according to the target image, a target parameter corresponding to the sample to be tested on the current image channel of the target image; the target parameter is used to indicate whether the sample to be tested meets a base balance condition on the current image channel;

[0043] A first obtaining unit is used to obtain a normalized parameter of the target image on the current image channel at least according to a brightness value of a pixel of the target image on the current image channel, when the target parameter indicates that the sample to be tested satisfies the base balance condition on the current image channel;

[0044] A second obtaining unit is used to obtain a normalized parameter of the target image on the current image channel according to the brightness values ​​of the pixels of the target image on all image channels when the target parameter indicates that the sample to be tested does not satisfy the base balance condition on the current image channel;

[0045] The normalization unit is used to normalize the brightness values ​​of the pixels on the current image channel according to the normalization parameters.

[0046] An electronic device, comprising:

[0047] A memory, used to store a computer program and data generated by the execution of the computer program;

[0048] A processor is used to execute the computer program to achieve: obtaining a target image, the target image is an image obtained by image acquisition of a sample to be tested under laser irradiation, the sample to be tested being dyed with a fluorescent dye; obtaining a target parameter corresponding to the sample to be tested on a current image channel of the target image according to the target image; the target parameter is used to characterize whether the sample to be tested satisfies a base balance condition on the current image channel; when the target parameter characterizes that the sample to be tested satisfies the base balance condition on the current image channel, obtaining a normalization parameter of the target image on the current image channel at least according to a pixel brightness value of the target image on the current image channel; when the target parameter characterizes that the sample to be tested does not satisfy the base balance condition on the current image channel, obtaining a normalization parameter of the target image on the current image channel according to the pixel brightness values ​​of the target image on all image channels; and normalizing the pixel brightness values ​​on the current image channel according to the normalization parameter.

[0049] It can be seen from the above technical scheme that in a method, device and electronic device for normalizing image brightness disclosed in the present application, after obtaining a target image of a sample to be tested under laser irradiation and dyed with a fluorescent material, a target parameter corresponding to the sample to be tested on the current image channel of the target image is obtained according to the target image; the target parameter represents whether the sample to be tested satisfies the base balance condition on the current image channel, based on this, when the target parameter represents that the sample to be tested satisfies the base balance condition on the current image channel, at least according to the pixel brightness value of the target image on the current image channel, the normalized parameter of the target image on the current image channel is obtained, and when the target parameter represents that the sample to be tested does not satisfy the base balance condition on the current image channel, the normalized parameter of the target image on the current image channel is obtained according to the pixel brightness value of the target image on all image channels; finally, the pixel brightness value on the current image channel is normalized according to the normalized parameter. It can be seen that in the present application, different methods are used to perform brightness normalization processing for base balanced and unbalanced image channels. For base balanced image channels, only the brightness value of the pixels on the current channel can be used for normalization processing, but for base unbalanced image channels, in addition to using the brightness value of the pixels on the current image channel, the brightness values ​​of the pixels on other image channels will also be combined for normalization processing, thereby avoiding the situation where the normalized brightness value is artificially high due to the use of only the brightness value of the pixels in the current channel due to base imbalance, thereby improving the accuracy of image brightness normalization processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0051] Figure 1 A flowchart of a method for normalizing image brightness provided in Embodiment 1 of the present application;

[0052] Figure 2 and Figure 3 They are respectively partial flow charts of a method for normalizing image brightness provided in Embodiment 1 of the present application;

[0053] Figure 4 A schematic diagram of the structure of a device for normalizing image brightness provided in Embodiment 2 of the present application;

[0054] Figure 5 Another structural schematic diagram of a device for normalizing image brightness provided in the second embodiment of the present application;

[0055] Figure 6 A schematic diagram of the structure of an electronic device provided in Embodiment 3 of the present application;

[0056] Figure 7 This is a processing flow chart for image brightness normalization in this application;

[0057] Figure 8 This is a comparison example of the brightness of the AG channel after normalization and normalization;

[0058] Fig. 9 This is a comparison example of the unnormalized cross-brightness map of all channels and the normalized brightness map of the original algorithm;

[0059] Fig.10 It is a schematic diagram of the ratio values ​​of the first 50 cycles of the four channels;

[0060] Fig.11 It is the ratio histogram of P99 and sumP99 of all channels;

[0061] Fig.12 It is a comparison diagram of the original brightness of each channel that is not normalized and the brightness of each channel after normalization using the solution of the present application;

[0062] Fig.13 This is a schematic diagram showing the comparison of results for different data using the normalization method in this application. DETAILED DESCRIPTION

[0063] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0064] refer to Figure 1 As shown, it is a flowchart of an implementation method of a normalization processing method of image brightness provided in the first embodiment of the present application, and the method can be applied to electronic devices capable of image processing, such as computers or servers. The technical solution in this embodiment is mainly used to improve the accuracy of image brightness normalization processing, so as to improve the accuracy of base recognition in the image.

[0065] Specifically, the method in this embodiment may include the following steps:

[0066] Step 101: Obtain a target image, where the target image is an image acquired by capturing an image of a sample to be tested under laser irradiation, wherein the sample to be tested is dyed with a fluorescent dye.

[0067] The sample to be tested is a sample that needs to be identified by base sequence. In this embodiment, the sample to be tested, which is irradiated by laser and dyed by fluorescent material, can be imaged by a base sequencing device to obtain a target image, which includes multiple image channels, such as A, C, G, T and other channels.

[0068] Step 102: According to the target image, the target parameters corresponding to the sample to be tested on the current image channel of the target image are obtained.

[0069] The target parameter is used to characterize whether the sample to be tested meets the base balance condition on the current image channel. The base balance condition refers to the condition that multiple bases are distributed evenly on the image channel.

[0070] Specifically, in this embodiment, at least the pixel values ​​of the pixels of the target image on the current image channel may be analyzed to obtain the target parameters.

[0071] Step 103: Determine whether the target parameter indicates that the sample to be tested meets the base balance condition on the current image channel. If the target parameter indicates that the sample to be tested does not meet the base balance condition on the current image channel, execute step 104; if the target parameter indicates that the sample to be tested meets the base balance condition on the current image channel, execute step 105.

[0072] Step 104: Obtain normalization parameters of the target image in the current image channel according to the brightness values ​​of the pixels of the target image in all image channels.

[0073] Specifically, step 104 can be implemented in the following manner:

[0074] First, the pixels of the target image on each image channel are sorted according to the brightness value; for example, in this embodiment, the pixels on each image channel are sorted according to the rule of brightness value from small to large;

[0075] Then, the brightness values ​​of the pixels ranked in the first percentile on all image channels are added to obtain a first value; the first percentile may be the 99th percentile, and the pixels ranked in the first percentile are: pixels ranked in the 99th percentile (denoted as P99); for example, in this embodiment, the brightness values ​​of the pixels ranked in the 99th percentile from the smallest to the largest brightness values ​​on all image channels are added to obtain a first value, which is denoted as SumP99;

[0076] Furthermore, the brightness values ​​of the pixels ranked in the second quantile on all image channels are added to obtain a second value; the second quantile may be the 30th quantile, and the pixels ranked in the second quantile are: pixels ranked in 30% (i.e., P30); for example, in this embodiment, the brightness values ​​of the pixels ranked in 30% from small to large on all image channels are added to obtain a second value, which is recorded as SumP30;

[0077] Based on this, the first value and the second value constitute a normalization parameter of the target image on the current image channel.

[0078] Furthermore, in step 104, before sorting the pixel points of the target image on each image channel according to the brightness values, the brightness values ​​of the pixel points of the target image on each image channel may be scaled according to the corresponding scaling dimension value on each image channel.

[0079] The corresponding scaling dimension value on the image channel can be obtained in advance based on multiple frames of sample images obtained by the base sequencing device to which the target image belongs in multiple sequencing cycles. For the same base sequencing device, the corresponding scaling dimension values ​​on different image channels may be different; and for different base sequencing devices, the corresponding scaling dimension values ​​on the same image channel may be different.

[0080] Step 105: obtaining a normalization parameter of the target image on the current image channel at least according to the brightness value of the pixels of the target image on the current image channel.

[0081] Specifically, step 105 can be implemented in the following manner:

[0082] First, the pixels of the target image on each image channel are sorted according to the brightness value; for example, in this embodiment, the pixels on each image channel are sorted according to the rule of brightness value from small to large;

[0083] Then, the brightness values ​​of the pixels ranked in the first percentile on the current image channel are added to obtain a third value; for example, in this embodiment, the brightness values ​​of the pixels ranked in 99% from small to large on the current image channel such as the AG channel are added to obtain a third value, which is recorded as P99;

[0084] Furthermore, the brightness values ​​of the pixels in the second percentile on all image channels are added to obtain a fourth value; for example, in this embodiment, the brightness values ​​of the pixels in the 30% percentile on all image channels are added to obtain a second value, which is recorded as SumP30;

[0085] The third value and the fourth value constitute the normalization parameter of the target image on the current image channel.

[0086] Furthermore, in step 105, before sorting the pixels of the target image on each image channel according to the brightness values, the brightness values ​​of the pixels of the target image on each image channel may be scaled according to the corresponding scaling dimension value on each image channel.

[0087] Step 106: normalize the brightness values ​​of the pixels on the current image channel according to the normalization parameters.

[0088] For example, when the target parameter indicates that the sample to be tested does not meet the base balance condition on the current image channel, SumP99 and SumP30 are used to normalize the brightness values ​​of the pixels on the current image channel; when the target parameter indicates that the sample to be tested meets the base balance condition on the current image channel, P99 and SumP30 are used to normalize the brightness values ​​of the pixels on the current image channel.

[0089] It can be seen from the above technical scheme that in a method, device and electronic device for normalizing image brightness provided in the first embodiment of the present application, after obtaining a target image of a sample to be tested under laser irradiation and dyed with a fluorescent material, a target parameter corresponding to the sample to be tested on the current image channel of the target image is obtained according to the target image; the target parameter represents whether the sample to be tested satisfies the base balance condition on the current image channel, based on this, when the target parameter represents that the sample to be tested satisfies the base balance condition on the current image channel, at least according to the pixel brightness value of the target image on the current image channel, the normalized parameter of the target image on the current image channel is obtained, and when the target parameter represents that the sample to be tested does not satisfy the base balance condition on the current image channel, the normalized parameter of the target image on the current image channel is obtained according to the pixel brightness value of the target image on all image channels; finally, the pixel brightness value on the current image channel is normalized according to the normalized parameter. It can be seen that in the present application, different methods are used to perform brightness normalization processing for base balanced and unbalanced image channels. For base balanced image channels, only the brightness value of the pixels on the current channel can be used for normalization processing, but for base unbalanced image channels, in addition to using the brightness value of the pixels on the current image channel, the brightness values ​​of the pixels on other image channels will also be combined for normalization processing, thereby avoiding the situation where the normalized brightness value is artificially high due to the use of only the brightness value of the pixels in the current channel due to base imbalance, thereby improving the accuracy of image brightness normalization processing.

[0090] In one implementation, the scaling dimension value corresponding to the target image on any target image channel is obtained by: Figure 2 As shown in:

[0091] Step 201: obtaining multiple frames of sample images, where the sample images are images corresponding to a base sequencing device to which a target image belongs in multiple sequencing cycles.

[0092] In this embodiment, for the base sequencing device to which the target image belongs, a sample image corresponding to each sequencing cycle is obtained, and the sample image is an image obtained by collecting the target sample under laser irradiation, and the target sample is stained with a fluorescent dye. The target sample and the sample to be tested in the above text can be the same sample.

[0093] Step 202: Sort the pixels of the sample image on each image channel according to the brightness value.

[0094] For example, in this embodiment, the pixels on each image channel are sorted according to the rule of brightness value from small to large.

[0095] Step 203: Add the brightness values ​​of the pixels on the target image channel that are sorted between the first quantile and the second quantile to obtain a fifth value.

[0096] For example, in this embodiment, the brightness values ​​of the pixels whose brightness values ​​are ranked from small to large and are between 30% and 99% on the target image channel, such as the AG channel, of each frame of the sample image are added to obtain the fifth value.

[0097] Step 204: Add the brightness values ​​of the pixels sorted between the first quantile and the second quantile on all image channels to obtain a sixth value.

[0098] For example, in this embodiment, the brightness values ​​of the pixels whose brightness values ​​are ranked from small to large and are between 30% and 99% on all image channels of each frame of sample image are added to obtain a sixth value.

[0099] Step 205: For each frame of sample image, obtain a ratio between the fifth value and the sixth value.

[0100] For example, for the target image channel of each frame of sample image, the fifth value is divided by the sixth value to obtain a ratio value of each frame of sample image on the target image channel.

[0101] Step 206: According to the number of sample images, the scale values ​​are averaged to obtain the corresponding scaling dimension value on the target image channel.

[0102] It should be noted that the multiple ratio values ​​corresponding to the multiple frames of sample images meet the ratio stability condition. For example, the variance between the ratio values ​​corresponding to each frame of sample images on the target image channel is lower than the variance threshold, that is, the ratio value corresponding to each frame of sample images on the target image channel is relatively stable, and there is no abnormally large or abnormally small ratio value.

[0103] In one implementation, in step 102, when obtaining the target parameters corresponding to the sample to be tested on the current image channel of the target image according to the target image, it can be implemented in the following manner: Figure 3 As shown in:

[0104] Step 301: Sort the pixels of the target image on each image channel according to the brightness value.

[0105] For example, in this embodiment, the pixel points of the target image on each image channel are sorted according to the rule of increasing brightness value.

[0106] Step 302: Add the brightness values ​​of the pixels of the target image that are sorted in the first quantile on the current image channel to obtain a seventh value.

[0107] For example, in this embodiment, the brightness values ​​of the pixels in the target image that are ranked in ascending order of brightness values ​​at 99% on the current image channel are added to obtain the seventh value.

[0108] Step 303: Add the brightness values ​​of the pixels in the first quantile on all image channels to obtain an eighth value.

[0109] For example, in this embodiment, the brightness values ​​of the pixels in the 99th percentile, which are sorted from small to large, on all image channels are added together to obtain the eighth value.

[0110] Step 304: at least the seventh value and the eighth value are used to obtain the target parameter corresponding to the sample to be tested on the current image channel.

[0111] Specifically, in step 304, when the ratio of the seventh value to the eighth value is greater than or equal to the balance threshold, a target parameter characterizing that the sample to be tested satisfies the base balance condition on the current image channel can be obtained; when the ratio of the seventh value to the eighth value is less than the balance threshold, a target parameter characterizing that the sample to be tested does not satisfy the base balance condition on the current image channel can be obtained.

[0112] For example, the balance threshold can be set to 0.8. When the ratio between the seventh value and the eighth value is less than 0.8, the target parameter indicates that the base balance condition is not met on the current image channel. When the ratio between the seventh value and the eighth value is greater than or equal to 0.8, the target parameter indicates that the base balance condition is met on the current image channel.

[0113] refer to Figure 4, is a schematic diagram of a normalized image brightness processing device provided in Embodiment 2 of the present application, which can be configured in an electronic device capable of image processing, such as a computer or a server. The technical solution in this embodiment is mainly used to improve the accuracy of image brightness normalization processing, so as to improve the accuracy of base recognition in the image.

[0114] Specifically, the device in this embodiment may include the following units:

[0115] An image acquisition unit 401 is used to acquire a target image, wherein the target image is an image acquired by acquiring an image of a sample to be tested under laser irradiation, wherein the sample to be tested is stained with a fluorescent dye;

[0116] The parameter obtaining unit 402 is used to obtain the target parameter corresponding to the sample to be tested on the current image channel of the target image according to the target image; the target parameter is used to indicate whether the sample to be tested meets the base balance condition on the current image channel;

[0117] A first obtaining unit 403 is used to obtain a normalized parameter of the target image on the current image channel at least according to a pixel brightness value of the target image on the current image channel, when the target parameter indicates that the sample to be tested satisfies the base balance condition on the current image channel;

[0118] A second obtaining unit 404 is used to obtain a normalized parameter of the target image on the current image channel according to the brightness values ​​of the pixels of the target image on all image channels when the target parameter indicates that the sample to be tested does not satisfy the base balance condition on the current image channel;

[0119] The normalization unit 405 is used to normalize the brightness values ​​of the pixels on the current image channel according to the normalization parameters.

[0120] It can be seen from the above technical scheme that in a normalization processing device for image brightness provided in Example 2 of the present application, after obtaining a target image of a sample to be tested under laser irradiation and dyed with a fluorescent material, a target parameter corresponding to the sample to be tested on the current image channel of the target image is obtained according to the target image; the target parameter represents whether the sample to be tested satisfies the base balance condition on the current image channel, based on this, when the target parameter represents that the sample to be tested satisfies the base balance condition on the current image channel, at least according to the pixel brightness value of the target image on the current image channel, the normalization parameter of the target image on the current image channel is obtained, and when the target parameter represents that the sample to be tested does not satisfy the base balance condition on the current image channel, the normalization parameter of the target image on the current image channel is obtained according to the pixel brightness value of the target image on all image channels; finally, the pixel brightness value on the current image channel is normalized according to the normalization parameter. It can be seen that in this embodiment, different methods are used to normalize the brightness of base-balanced and unbalanced image channels. For base-balanced image channels, only the brightness value of the pixels on the current channel can be used for normalization. However, for base-unbalanced image channels, in addition to using the brightness value of the pixels on the current image channel, the brightness values ​​of the pixels on other image channels are also used for normalization. This avoids the situation where the normalized brightness value is artificially high due to the use of only the brightness value of the pixels in the current channel due to base imbalance, thereby improving the accuracy of image brightness normalization processing.

[0121] In one implementation, the second obtaining unit 404 is specifically used to: sort the pixel points of the target image on each image channel according to the brightness value; add the brightness values ​​of the pixel points sorted in the first quartile on all the image channels to obtain a first value; add the brightness values ​​of the pixel points sorted in the second quartile on all the image channels to obtain a second value; wherein the first value and the second value constitute the normalization parameter of the target image on the current image channel.

[0122] In one implementation, the first obtaining unit 403 is specifically used to: sort the pixel points of the target image on each image channel according to the brightness value; add the brightness values ​​of the pixel points sorted in the first quartile on the current image channel to obtain a third value; add the brightness values ​​of the pixel points sorted in the second quartile on all the image channels to obtain a fourth value; wherein the third value and the fourth value constitute the normalization parameter of the target image on the current image channel.

[0123] Furthermore, before sorting the pixel points of the target image on each image channel according to the brightness values, the first acquisition unit 403 or the second acquisition unit 404 is also used to: scale the brightness values ​​of the pixel points of the target image on each image channel according to the corresponding scaling dimension value on each image channel.

[0124] Preferably, the device in this embodiment may also include the following units: Figure 5 As shown in:

[0125] A dimension obtaining unit 406 is used to obtain a scaling dimension value corresponding to a target image on any target image channel;

[0126] Among them, the dimension acquisition unit 406 is specifically used to: obtain multiple frames of sample images, where the sample images are images corresponding to the base sequencing device to which the target image belongs in multiple sequencing cycles; sort the pixels of the sample images on each of the image channels according to the brightness values; add the brightness values ​​of the pixels sorted between the first quartile and the second quartile on the target image channel to obtain a fifth value; add the brightness values ​​of the pixels sorted between the first quartile and the second quartile on all the image channels to obtain a sixth value; obtain the ratio value between the fifth value and the sixth value for each frame of the sample image; and average the ratio values ​​according to the number of the sample images to obtain the corresponding scaled dimension value on the target image channel.

[0127] Preferably, the multiple ratio values ​​corresponding to the multiple frames of sample images satisfy a ratio stability condition.

[0128] In one implementation, the parameter acquisition unit 402 is specifically used to: sort the pixel points of the target image on each image channel according to the brightness value; add the brightness values ​​of the pixel points of the target image sorted in the first quartile on the current image channel to obtain a seventh value; add the brightness values ​​of the pixel points sorted in the first quartile on all the image channels to obtain an eighth value; at least the seventh value and the eighth value are used to obtain the target parameter corresponding to the sample to be tested on the current image channel, for example: when the ratio of the seventh value to the eighth value is greater than or equal to the balance threshold, the target parameter characterizing that the sample to be tested meets the base balance condition on the current image channel is obtained; when the ratio of the seventh value to the eighth value is less than the balance threshold, the target parameter characterizing that the sample to be tested does not meet the base balance condition on the current image channel is obtained.

[0129] It should be noted that the specific implementation of each unit in this embodiment can refer to the corresponding content in the previous text and will not be described in detail here.

[0130] refer to Figure 6 , is a schematic diagram of the structure of an electronic device provided in Embodiment 3 of the present application, and the electronic device may include the following structure:

[0131] Memory 601, used to store computer programs and data generated by the execution of the computer programs;

[0132] The processor 602 is used to execute the computer program to achieve: obtaining a target image, the target image being an image obtained by performing image acquisition on a sample to be tested under laser irradiation, the sample to be tested being stained with a fluorescent dye; obtaining a target parameter corresponding to the sample to be tested on a current image channel of the target image according to the target image; the target parameter is used to characterize whether the sample to be tested satisfies a base balance condition on the current image channel; when the target parameter characterizes that the sample to be tested satisfies the base balance condition on the current image channel, obtaining a normalization parameter of the target image on the current image channel at least according to a pixel brightness value of the target image on the current image channel; when the target parameter characterizes that the sample to be tested does not satisfy the base balance condition on the current image channel, obtaining a normalization parameter of the target image on the current image channel according to the pixel brightness values ​​of the target image on all image channels; and normalizing the pixel brightness values ​​on the current image channel according to the normalization parameter.

[0133] It can be seen from the above technical scheme that in an electronic device provided in Example 3 of the present application, after obtaining a target image of a sample to be tested under laser irradiation and dyed with a fluorescent material, a target parameter corresponding to the sample to be tested on the current image channel of the target image is obtained according to the target image; the target parameter represents whether the sample to be tested satisfies the base balance condition on the current image channel, based on this, when the target parameter represents that the sample to be tested satisfies the base balance condition on the current image channel, at least according to the pixel brightness value of the target image on the current image channel, the normalized parameter of the target image on the current image channel is obtained, and when the target parameter represents that the sample to be tested does not satisfy the base balance condition on the current image channel, the normalized parameter of the target image on the current image channel is obtained according to the pixel brightness value of the target image on all image channels; finally, the pixel brightness value on the current image channel is normalized according to the normalized parameter. It can be seen that in this embodiment, different methods are used to normalize the brightness of base-balanced and unbalanced image channels. For base-balanced image channels, only the brightness value of the pixels on the current channel can be used for normalization. However, for base-unbalanced image channels, in addition to using the brightness value of the pixels on the current image channel, the brightness values ​​of the pixels on other image channels are also used for normalization. This avoids the situation where the normalized brightness value is artificially high due to the use of only the brightness value of the pixels in the current channel due to base imbalance, thereby improving the accuracy of image brightness normalization processing.

[0134] In combination with the above embodiments, the present application provides an image brightness normalization processing solution based on inter-channels. The algorithm provided by the present application has strong compatibility with base balanced and unbalanced sequencing, especially for base unbalanced samples, the accuracy is improved by 10%-90% compared with the previous normalization algorithm.

[0135] like Figure 7 As shown in , the specific steps of this application are as follows:

[0136] First, the prior parameters are counted: for each machine (base sequencing device), the original alignment algorithm is used to count the P99-P30 ratios between different channels of the base-balanced samples. The ratio of each channel is: the original brightness P99-P30 value of this channel divided by the sum of the original brightness P99-P30 values ​​of all channels.

[0137] Each channel in each cycle will have a ratio. As the number of cycles increases, the ratio between channels should remain unchanged, because the efficiency of laser excitation of a channel is always relatively constant. Then we can calculate the ratio of each cycle, calculate the average value and save a ratio value for each channel, and then save it as an empirical dimension. The ratio values ​​of each machine are different and cannot be mixed.

[0138] In addition, if the ratio between each cycle is unstable, this machine cannot perform experiments on base imbalance.

[0139] Then, in practical applications, whether it is base-balanced or base-unbalanced data, firstly, the rawint (original signal value) of each channel must be divided by the corresponding dimension to scale the values ​​of each channel.

[0140] After that, the brightness values ​​of all channels are combined: the total brightness of P99 and P30 is summed up, which can be referred to as SumP99 and SumP30. In theory, SumP99 can be used as the signal value of all channels, but in practice, it is found that for base-balanced data, the result of using SumP99 is more than 2% worse than the result of using each channel's P99, and the selection of P30 is not much different, that is, SumP99 is not completely compatible with base-balanced and base-unbalanced situations.

[0141] Based on this, the present application distinguishes between these two situations and needs to determine whether the current cycle is a base-balanced situation. Therefore, the brightness of each channel is sorted separately to obtain the corresponding brightness of P99 and P30.

[0142] in:

[0143] If the ratio of P99 / SumP99 of the current channel is less than 0.8, it is judged as base imbalance, and if it is greater than or equal to 0.8, it is judged as base balance. The value of 0.8 is obtained through statistics and can distinguish the two situations.

[0144] Specifically, when the bases are balanced, the normalization uses the P99 of the current channel, and P30 uses SumP30; when the bases are unbalanced, the normalization uses the sum of P99 and P30 of all channels, that is, SumP99 and SumP30.

[0145] It can be seen that in the present application, the image with very dim luminescence due to base deletion can be normalized, and the information of other channels can be used to obtain the values ​​of P99 and P30, and the base balance and base imbalance can be effectively judged, so that this algorithm can be compatible with both sequencing without using two sets of algorithms.

[0146] The following is an example of the technical solution of this application:

[0147] When the content of each base in the fluorescent molecular group is greater than 12.5% ​​(non-base imbalance) and the light is emitted at random positions, the original brightness is uneven due to the different excitation efficiencies of the fluorescent molecular groups in each channel. Figure 8The comparison example of the brightness of the AG channel after normalization and before normalization is shown in the figure. The brightness of the AG channel reaches 2500, but the brightness of the G channel only reaches 1500. The 100-1500 of the G channel represents that the G base in the fluorescent molecular group is excited, while the 1000-1500 of the A channel is still at a medium-low level. If the base recognition is performed directly according to the original brightness value, the brightness of the A channel will be mostly higher than the G base, which is incorrect. Therefore, the original signal brightness needs to be normalized to the same scale before the comparison between channels can be performed.

[0148] The normalization method currently used is percentile normalization. Take percentile 30 and percentile 99 for normalization, and normalize each channel separately. The normalization formula is shown in formula (1):

[0149] I norG =(IP 30 ) / (P 99 -P 30 ) (1)

[0150] Where, I is the original brightness, I norG is the normalized brightness value. When the content of a certain base in the fluorescent molecule group is less than 12.5% ​​(base imbalance) and the random position emits light, using the original algorithm will lead to serious errors.

[0151] like Fig. 9 As shown in the figure, it is a comparison example of the cross-brightness map of all channels that are not normalized and the brightness map after normalization by the original algorithm. In the original signal brightness map, it can be seen that the brightness of channel A is very low compared to other brightnesses, only a few hundred, and the laser excitation is normal, which means that there is only background, no A base luminescence, which means that the A base content of the current cycle is very low. Later it was proved that no A base was added when the machine was put on. In such an extreme case, it is unreasonable to use the original normalization algorithm to forcibly increase the brightness of channel A to the same size as other channels. This will mistake the background of channel A as a signal, causing serious misjudgment of the results.

[0152] Based on this, the present application proposes a normalization method based on channels.

[0153] (1) Statistical prior parameters.

[0154] For each machine, the original registration algorithm is used to calculate the ratio of P99-P30 between different channels of the base balance sample (Ecoli). The ratio of each channel is: the original brightness P99-P30 value of this channel / the sum of the original brightness P99-P30 values ​​of all channels. Each channel of each cycle will have a ratio. In theory, as the cycle increases, the ratio between channels should remain unchanged, because the efficiency of laser excitation of a certain channel should be relatively constant. Then the mean ratio of each cycle can be calculated, and finally a ratio value for each channel can be saved, and then saved as an empirical dimension. The ratio values ​​of each machine are different and cannot be mixed. Fig.10 The figure shows the ratio values ​​of the first 50 cycles of the four channels. Finally, the first five FOVs are removed and the average of all cycles is calculated to obtain four ratio values, 0.44, 0.19, 0.17, and 0.18.

[0155] (2) Removing dimensions: Regardless of whether the data is base-balanced or base-unbalanced, we must first divide the rawint (original signal value) of each channel by the corresponding dimension to scale the values ​​of each channel.

[0156] (3) Combine the brightness values ​​of all channels and calculate the total brightness of P99 and P30, which can be referred to as SumP99 and SumP30. In theory, SumP99 can be used as the signal value of all channels, but in practice, it is found that for base-balanced data, the result of using SumP99 is more than 2% worse than the result of using the P99 of each channel, and the selection of P30 is not much different, that is, SumP99 is not completely compatible with base-balanced and base-unbalanced situations. Therefore, it is necessary to distinguish between the two situations.

[0157] (4) Determine whether it is a base imbalance situation. Therefore, the brightness of each channel is sorted separately to obtain the corresponding brightness of P99 and P30. Fig.11 As shown in the figure, it is the ratio histogram of P99 and sumP99 of all channels. 0.8 can capture all the cases of base imbalance (red). Finally, if the ratio of P99 / SumP99 of the current channel is less than 0.8, it is judged as base imbalance. This value is obtained through statistics and can distinguish the two cases. If the ratio of P99 / SumP99 of the current channel is less than 0.8, it is judged as base imbalance, otherwise it is base balance.

[0158] (5) Select the normalization method. If the bases are balanced, the normalization uses the P99 of the current channel, and P30 uses SumP30. If the bases are unbalanced, the normalization uses the sum of P99 and P30 of all channels, that is, SumP99 and SumP30.

[0159] (6) Normalization, still using formula (1), using the selected P99.

[0160] (7) In the case of base imbalance, the effect of the new normalization method is as follows Fig.12 As shown in the figure, it is a comparison of the original brightness of each channel without normalization and the brightness of each channel after normalization. The new normalization method maintains the original brightness shape, and the channel with low brightness still maintains a low brightness state after normalization. Finally, both base-balanced and unbalanced data were tested, and it was found that both cases are compatible. The splitting rate of base-balanced data decreased by 0.7% compared with the original algorithm, and the splitting rate of base-unbalanced data was significantly improved, increasing by 70%-90%. Fig.13 As shown in , this is a schematic diagram comparing the results of the new normalization method for different data.

[0161] The main purpose of this application is to provide a normalization method that is compatible with extremely sparse fluorescent molecules, so that the brightness of channels with low base content is always at a low level. Solve the limitations of BGI sequencers for base imbalance sequencing. In addition, this application can replace this normalization method by using a deep learning method.

[0162] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0163] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0164] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented directly using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0165] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for normalizing image brightness, characterized in that: The method comprises: Obtaining a target image, wherein the target image is an image obtained by collecting an image of a sample to be tested under laser irradiation, wherein the sample to be tested is stained with a fluorescent dye; According to the target image, obtaining a target parameter corresponding to the sample to be tested on the current image channel of the target image; the target parameter is used to indicate whether the sample to be tested meets a base balance condition on the current image channel; When the target parameter represents that the sample to be tested satisfies the base balance condition on the current image channel, obtaining a normalized parameter of the target image on the current image channel at least according to the brightness value of the pixel of the target image on the current image channel; When the target parameter indicates that the sample to be tested does not satisfy the base balance condition on the current image channel, obtaining a normalized parameter of the target image on the current image channel according to the brightness values ​​of the pixels of the target image on all image channels; According to the normalization parameter, the brightness values ​​of the pixels on the current image channel are normalized.

2. The method according to claim 1, characterized in that Obtaining normalized parameters of the target image on the current image channel according to pixel brightness values ​​of the target image on all image channels, including: Sort the pixels of the target image on each image channel according to the brightness value respectively; Adding the brightness values ​​of the pixels in the first quantile on all the image channels to obtain a first value; Adding the brightness values ​​of the pixels in the second quantile on all the image channels to obtain a second value; The first value and the second value constitute a normalization parameter of the target image on the current image channel.

3. The method according to claim 1, characterized in that Obtaining normalized parameters of the target image on the current image channel according to the brightness values ​​of pixels of the target image on the current image channel includes: Sort the pixels of the target image on each image channel according to the brightness value respectively; Add the brightness values ​​of the pixels in the first quantile on the current image channel to obtain a third value; Adding the brightness values ​​of the pixels in the second quantile on all the image channels to obtain a fourth value; The third value and the fourth value constitute a normalization parameter of the target image on the current image channel.

4. The method according to claim 2 or 3, characterized in that: Before sorting the pixels of the target image on each image channel according to the brightness values, the method further includes: According to the corresponding scaling dimension value on each image channel, the brightness value of the pixel point of the target image on each image channel is scaled respectively.

5. The method according to claim 4, characterized in that The scaling dimension value corresponding to the target image on any target image channel is obtained in the following manner: Obtaining multiple frames of sample images, where the sample images are images corresponding to the base sequencing device to which the target image belongs in multiple sequencing cycles; Sort the pixel points of the sample image on each image channel according to the brightness value respectively; Adding the brightness values ​​of the pixels on the target image channel that are sorted between the first quantile and the second quantile to obtain a fifth value; Adding the brightness values ​​of the pixels on all the image channels that are sorted between the first quantile and the second quantile to obtain a sixth value; For each frame of the sample image, respectively, obtaining a ratio between the fifth value and the sixth value; The scale values ​​are averaged according to the number of the sample images to obtain the corresponding scaling dimension value on the target image channel.

6. The method according to claim 5, characterized in that The multiple ratio values ​​corresponding to the multiple frames of sample images meet a ratio stability condition.

7. The method according to claim 1, 2 or 3, characterized in that: According to the target image, obtaining target parameters corresponding to the sample to be tested on the current image channel of the target image includes: Sort the pixels of the target image on each image channel according to the brightness value respectively; Adding the brightness values ​​of the pixels of the target image that are sorted in the first quantile on the current image channel to obtain a seventh value; Adding the brightness values ​​of the pixels in the first quantile on all the image channels to obtain an eighth value; At least the seventh value and the eighth value are used to obtain the target parameter corresponding to the sample to be tested on the current image channel.

8. The method according to claim 7, characterized in that At least the seventh value and the eighth value are used to obtain the target parameter corresponding to the sample to be tested on the current image channel, including: When the ratio of the seventh value to the eighth value is greater than or equal to the balance threshold, obtaining a target parameter characterizing that the sample to be tested satisfies a base balance condition on the current image channel; When the ratio of the seventh value to the eighth value is smaller than the balance threshold, a target parameter is obtained that indicates that the sample to be tested does not satisfy a base balance condition on the current image channel.

9. A device for normalizing image brightness, characterized in that: The device comprises: An image acquisition unit, used to acquire a target image, wherein the target image is an image acquired by acquiring an image of a sample to be tested under laser irradiation, wherein the sample to be tested is dyed with a fluorescent dye; A parameter obtaining unit, used for obtaining, according to the target image, a target parameter corresponding to the sample to be tested on the current image channel of the target image; the target parameter is used to indicate whether the sample to be tested meets a base balance condition on the current image channel; A first obtaining unit is used to obtain a normalized parameter of the target image on the current image channel at least according to a brightness value of a pixel of the target image on the current image channel, when the target parameter indicates that the sample to be tested satisfies the base balance condition on the current image channel; A second obtaining unit is used to obtain a normalized parameter of the target image on the current image channel according to the brightness values ​​of the pixels of the target image on all image channels when the target parameter indicates that the sample to be tested does not satisfy the base balance condition on the current image channel; The normalization unit is used to normalize the brightness values ​​of the pixels on the current image channel according to the normalization parameters.

10. An electronic device, characterized in that: include: A memory, used to store a computer program and data generated by the execution of the computer program; A processor is used to execute the computer program to achieve: obtaining a target image, wherein the target image is an image obtained by collecting an image of a sample to be tested under laser irradiation, wherein the sample to be tested is dyed with a fluorescent dye; and obtaining, according to the target image, a target parameter corresponding to the sample to be tested on a current image channel of the target image; The target parameter is used to characterize whether the sample to be tested meets the base balance condition on the current image channel; When the target parameter represents that the sample to be tested satisfies the base balance condition on the current image channel, obtaining a normalized parameter of the target image on the current image channel at least according to the brightness value of the pixel of the target image on the current image channel; When the target parameter represents that the sample to be tested does not satisfy the base balance condition on the current image channel, the normalization parameter of the target image on the current image channel is obtained according to the pixel brightness values ​​of the target image on all image channels; and the pixel brightness values ​​on the current image channel are normalized according to the normalization parameter.

Images