CN116309491A - Method for evaluating definition and point roundness of fluorescent image - Google Patents
Method for evaluating definition and point roundness of fluorescent image Download PDFInfo
- Publication number
- CN116309491A CN116309491A CN202310295465.3A CN202310295465A CN116309491A CN 116309491 A CN116309491 A CN 116309491A CN 202310295465 A CN202310295465 A CN 202310295465A CN 116309491 A CN116309491 A CN 116309491A
- Authority
- CN
- China
- Prior art keywords
- image
- vector
- definition
- stretching
- sub
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 43
- 108090000623 proteins and genes Proteins 0.000 claims abstract description 18
- 239000013598 vector Substances 0.000 claims description 128
- 238000010606 normalization Methods 0.000 claims description 23
- 238000010586 diagram Methods 0.000 claims description 17
- 238000002073 fluorescence micrograph Methods 0.000 claims description 15
- 238000000018 DNA microarray Methods 0.000 claims description 13
- 230000008859 change Effects 0.000 claims description 13
- 238000004458 analytical method Methods 0.000 claims description 10
- 238000004590 computer program Methods 0.000 claims description 8
- 238000003384 imaging method Methods 0.000 claims description 7
- 108020004707 nucleic acids Proteins 0.000 claims description 6
- 102000039446 nucleic acids Human genes 0.000 claims description 6
- 150000007523 nucleic acids Chemical class 0.000 claims description 6
- 238000007781 pre-processing Methods 0.000 claims description 5
- 230000006870 function Effects 0.000 claims description 3
- 238000012886 linear function Methods 0.000 claims description 3
- 238000012163 sequencing technique Methods 0.000 abstract description 46
- 238000006243 chemical reaction Methods 0.000 abstract description 13
- 230000009466 transformation Effects 0.000 description 9
- 230000008569 process Effects 0.000 description 8
- 108020004414 DNA Proteins 0.000 description 5
- 239000011159 matrix material Substances 0.000 description 5
- 102000053602 DNA Human genes 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 4
- 238000001514 detection method Methods 0.000 description 4
- 238000012165 high-throughput sequencing Methods 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 239000012634 fragment Substances 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 230000003321 amplification Effects 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000003199 nucleic acid amplification method Methods 0.000 description 1
- 239000002773 nucleotide Substances 0.000 description 1
- 125000003729 nucleotide group Chemical group 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
- 238000000844 transformation Methods 0.000 description 1
- 238000011426 transformation method Methods 0.000 description 1
- 230000001131 transforming effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30168—Image quality inspection
Landscapes
- Engineering & Computer Science (AREA)
- Quality & Reliability (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Investigating, Analyzing Materials By Fluorescence Or Luminescence (AREA)
Abstract
The invention relates to a method for evaluating the definition and the dot roundness of a fluorescent image, belonging to the field of gene sequencing. The method provided by the invention does not depend on any special mark on a gene sequencing chip, and can judge the definition of the image and the dot roundness only by using a fluorescent image in normal sequencing reaction. The operation steps are simple, and the method has strong robustness.
Description
Technical Field
The invention relates to a method for evaluating the definition and the dot roundness of a fluorescent image, belonging to the field of gene sequencing.
Background
During gene sequencing, large amounts of data are generated. Typically, second generation sequencing is the detection of bases using signals generated during chemical reactions. At the time of detection, the signal of the single DNA molecule is very weak and is difficult to detect, so that the single DNA molecule or fragment is replicated to a large number, e.g., 2, by amplification or the like 10 To facilitate detection of the signal. During the detection, each chemical reaction causes the complementary strand of the DNA strand to be detected to be extended by 1 base or more. Depending on the sequencing method, for example, the 3-terminal closed Illumina-like sequencing method extends 1 base at a time; for example, 2+2 type sequencingThe method may be extended by a plurality of bases at a time. Sequencing reactions that extend multiple bases at a time may also be referred to as multiple base sequencing reactions. In either method, in the sequencing process, multiple chemical reactions are required for the same DNA molecule or fragment to obtain information about the DNA to be tested having a certain length. During high-throughput sequencing, the sequencing chip moves relative to a picture acquisition component such as an objective lens. The high-flux sequencing chip has a large number of luminous sites, the luminous sites are round or approximately round under the condition of focusing an accurate image clearly, and the shape of the luminous sites in the obtained unclear image can be round, approximately round or oval when focusing is not performed. Focusing is needed when photographing the sequencing chip each time, so that the image is clear and is not blurred; therefore, it is necessary to develop an evaluation method for focusing and roundness of the image, to screen out a clear image by analyzing the focusing value and the roundness value, and to adjust the photographing scheme of the camera according to the focusing value and the roundness value, so as to photograph a clearer image. The invention discloses a method for evaluating image definition and point roundness by utilizing local maxima of a high-frequency region of a frequency domain image.
Disclosure of Invention
The invention discloses a method for evaluating the definition and the dot roundness of a fluorescent image, which is applied to a biochip and is characterized by comprising the following steps:
s1, acquiring a fluorescence image shot by the biochip at a preset shooting position;
s2, preprocessing the fluorescent image to obtain a preprocessed image;
s3, dividing the preprocessed image into N sub-areas, wherein N is more than or equal to 2; performing discrete Fourier transform on each sub-region to obtain a local maximum value set of the frequency domain image in a high frequency region;
s4, selecting non-centrosymmetric 4 high-frequency local maxima from the local maxima set, and converting the non-centrosymmetric 4 high-frequency local maxima into stretching vectors of corresponding pit images in a first direction and a second direction under a pixel coordinate system; the pit image is a site on a biochip for binding nucleic acid molecules and generating fluorescent molecules; the first direction and the second direction are perpendicular;
s5: selecting M sub-areas closest to the center of the fluorescent image, calculating the average value of stretching vectors of the M sub-areas in the first direction and the second direction, and calculating the distance of the average value of the stretching vectors, wherein the distance represents the point roundness of the fluorescent image;
s6: calculating the difference between the stretching vector of each sub-area in the first direction and the second direction and the average value of the stretching vectors of the M sub-areas in the first direction and the second direction to obtain a vector difference value, wherein the maximum length value of the vector difference value in the N sub-areas is the definition of the fluorescent image;
s7: comparing the point circularity obtained in the step S5 with a preset point circularity threshold value, and if the point circularity is smaller than the preset threshold value, judging that the image is not dithered, wherein the image is used for subsequent analysis; and (3) comparing the definition obtained in the step (S6) with a preset definition threshold, and if the definition is smaller than the preset threshold, judging that the image is clear, wherein the image is used for subsequent analysis.
According to a preferred embodiment, the surface of the biochip has a plurality of micro-pits, which are sites where fluorescence reaction occurs, and the pit period remains uniform after imaging; the pit period is the number of pixels that a single pit occupies after imaging.
According to a preferred embodiment, the preprocessing in S2 includes normalization operations including, but not limited to, linear function normalization, nonlinear function normalization, L2-norm normalization, zero mean normalization, and the like.
According to a preferred embodiment, the operation of converting the high frequency local maxima into stretching vectors in the first direction and the second direction in the pixel coordinate system in S4 comprises: under a pixel coordinate system, subtracting two high-frequency local maxima representing included angles of a first direction and a second direction to obtain an oblique stretching vector in the included angle direction of the first direction and the second direction, wherein the included angle between the oblique stretching vector and the first direction is 0-90 degrees; under a pixel coordinate system, two high-frequency local maxima representing the first direction or the second direction are subtracted to obtain a horizontal vertical vector of the first direction or the second direction, and an included angle between the horizontal vertical vector and the first direction is 0 degree or 90 degrees.
According to a preferred embodiment, the oblique stretching vector has an angle of 30 ° to 60 °, preferably 45 ° to 50 °, to the first direction.
According to a preferred embodiment, the distances described in S5 include, but are not limited to, euclidean distance, manhattan distance, chebyshev distance, min Shi distance, normalized euclidean distance, cosine similarity, mahalanobis distance, correlation distance, jaccard distance, and the like.
According to a preferred embodiment, the method further comprises:
s8: comparing the point circularity obtained in the step S5 with a preset point circularity threshold value, if the point circularity is larger than the threshold value, drawing a vector diagram according to the stretching vector average value obtained in the step S5, observing the change of the vector diagram, and correcting the photographing flow of the instrument according to the change; comparing the definition obtained in the step S6 with a preset definition threshold, if the definition is larger than the definition threshold, drawing a vector diagram according to N vector difference values obtained in the step S6, observing the change of the vector diagram, and correcting the photographing flow of the instrument according to the change.
The invention also discloses a gene sequencer which is characterized by comprising a processor, wherein the processor is used for realizing the method for evaluating the definition and the dot roundness of the fluorescent image when executing the computer program stored in the memory.
The invention also discloses a non-transitory computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when executed by a processor, implements a method of assessing fluorescence image sharpness and dot roundness as described above.
The beneficial effects of the invention are that
Compared with the prior art, the method for evaluating the definition and the dot roundness of the fluorescent image has the following advantages:
1. the method does not depend on any special mark on a gene sequencing chip, and can judge the definition of the image and the dot roundness only by using a fluorescent image in normal sequencing reaction. The operation steps are simple, and the method has strong robustness.
2. If the image is in virtual focus or the result of the point circularity is poor, a vector diagram can be drawn according to the mean vector and the difference vector in the calculation process, the shaking direction of the roundness difference of the image point and the degree of the virtual focus are judged, and the photographing scheme of the camera is corrected, so that a clearer image can be photographed conveniently.
Drawings
Fig. 1. Fig. 1 (a) shows that the pit image in the time domain image is biased to the anticlockwise 45 DEG direction, fig. 1 (b) shows that the corresponding frequency domain image is biased to the upper right corner point than the upper left corner point.
Fig. 2 (a) shows that the pit image in the time domain image is biased to the clockwise 45 DEG direction, and fig. 2 (b) shows that the corresponding frequency domain image is biased to the upper left corner point than the upper right corner point.
Fig. 3 (a) is an original image, and fig. 3 (b) is a histogram result of the original image, wherein the abscissa indicates brightness and the ordinate indicates the number of pixels.
Fig. 4. Fig. 4 (a) is a normalized image, and fig. 4 (b) is a histogram result of the normalized image.
Fig. 5. Feature points in the frequency domain image, wherein the middle point is a zero frequency point, the four points marked in the rest are respectively an upper left corner point, an upper edge point, an upper right corner point and a left edge point, and the four other points are respectively symmetrical with the four points about the center of the zero frequency point.
Fig. 6 is a vector diagram for correcting a photographing of a camera, in which the length of an arrow indicates the length of the vector, i.e., the magnitude of a focus value, and the direction of the arrow indicates the direction of the vector.
Detailed Description
In high throughput sequencing, a gene sequencing chip is typically used for sequencing. The features of the gene sequencing chip are described in applicant's previous patents. For example, in patent CN201910156574.0, example 1 or other parts of the specification, the structure of a gene sequencing chip is described. The interior surface of a typical gene sequencing chip has micro-pits or other possible structures for distinguishing each sequenced data point. The gene sequencing chip of the invention refers to a gene sequencing chip with or without micro pits.
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 this invention belongs. All patents, patent applications, published applications, papers, and other publications mentioned herein (above and below) are incorporated by reference in their entirety. If a definition and/or description set forth herein contradicts or is inconsistent with any definition set forth in the patents, patent applications, published applications and other publications incorporated herein by reference, the definition and/or description set forth herein controls.
The following description of various exemplary embodiments is merely exemplary and explanatory and is not to be construed as limiting or restrictive in any way. Other embodiments, features, objects, and advantageous aspects of the present teachings will be apparent from the description and drawings, and from the claims.
Interpretation of the terms
Discrete fourier transform (Discrete)
Fourier
Transform，DFT)
For a fast and efficient processing and analysis of the image, complex temporal or spatial signals may be transformed into a transform in the form of a frequency component representation structure, i.e. a frequency domain transform. The image frequency domain transformation mode includes various common modes including discrete fourier transformation, discrete wavelet transformation, discrete cosine transformation, fast fourier transformation and the like, and the frequency domain transformation mode can also be a combination method of various transformation methods, such as a combination method of firstly performing discrete wavelet transformation, then performing discrete cosine transformation and the like. In practical application, discrete fourier transform can be performed on the original sequencing image, so as to obtain a pixel matrix of the sequencing image in a frequency domain space. The discrete fourier transform, which is a form in which fourier transforms take on discrete form in both the time and frequency domains, transforms samples of a time domain signal into samples in the Discrete Time Fourier Transform (DTFT) frequency domain.
Frequency domain image
A frequency domain image or spectrogram, the frequency domain (frequency domain), whose argument is frequency, i.e. the horizontal axis is frequency, and the vertical axis is the amplitude of the frequency signal, i.e. the spectrogram. The spectrogram describes the frequency structure of the signal and the frequency versus the amplitude of the frequency signal.
Local maxima in high frequency region
The local maxima of the high frequency region represent the portions of the image that vary strongly, i.e. edges (contours) or noise and detail portions of the image. Mainly the measurement of the edges and contours of the image.
Image normalization
Image normalization refers to the process of transforming an image into a fixed standard form by performing a series of standard process transformations on the image, which standard image is referred to as a normalized image. The normalization processing is to adjust the feature values to a similar range. Normalization of the images is more beneficial to computer automatic analysis processing. Typically, normalization also performs the operation of subtracting the mean value divided by the variance, which removes the mean luminance value (intensity) of the image.
Distance of
Mathematical distance generally refers to a distance using two vectors, where the meaning of distance is consistent with the usual meaning, and common distance algorithms may include: euclidean distance, manhattan distance, chebyshev distance, min Shi distance, normalized euclidean distance, cosine similarity, mahalanobis distance, correlation distance, jekade distance, and the like.
Pixel coordinate system
A direct coordinate system u-v in pixels is established with the upper left corner of the image as the origin. The horizontal axis is u, and the right direction is the positive direction; the vertical axis is v, the downward direction is positive, u and v represent the number of rows and columns of pixels, and there is no physical unit.
Micro pit
The micro-pits are also called micro-reaction chambers, micro units, micro wells and the like, are micro-nano recessed structures arranged on the surface of the chip in an array manner, and are sites for fluorescence reaction. The micro-pits may take on a variety of shapes, for example: the shape is a cylinder, a truncated cone, a groove, a truncated cone-like shape, a hexagonal column-like structure, or a combination thereof. In the present invention, in a gene sequencing chip having micropits, the period of the micropits on the chip surface needs to be kept uniform, and the pit period refers to the number of pixels occupied by a single micropit after imaging.
Preset threshold value
In the present invention, the preset threshold may be an empirical value preset empirically by a person skilled in the art, and the threshold may be a fixed value.
Detailed Description
In the second generation sequencing technology, a plurality of sequencing companies represented by Illumina and Huada acquire high-throughput sequencing images, process the images and further acquire sequence information of nucleic acid to be detected. In the process of photographing or scanning the sequencing reaction, the imaging system is inevitably dithered, even if the imaging system is slightly dithered, the deviation of the sequencing image is caused, and the accuracy of the sequencing result is directly influenced by the definition of the sequencing image mainly in two aspects of definition and dot circle degree of the image, so that in order to ensure the high accuracy of the sequencing result, the high definition of the sequencing image needs to be ensured, namely, an effective method for evaluating the definition and dot circle degree of the sequencing image is required.
The invention provides a method for evaluating the definition and the dot roundness of a fluorescent image, which is applied to a biochip and is characterized by comprising the following steps:
s1, acquiring a fluorescence image shot by the biochip at a preset shooting position;
s2, preprocessing the fluorescent image to obtain a preprocessed image;
s3, dividing the preprocessed image into N sub-areas, wherein N is more than or equal to 2; performing discrete Fourier transform on each sub-region to obtain a local maximum value set of the frequency domain image in a high frequency region;
s4, selecting non-centrosymmetric 4 high-frequency local maxima from the local maxima set, and converting the non-centrosymmetric 4 high-frequency local maxima into stretching vectors of corresponding pit images in a first direction and a second direction under a pixel coordinate system; the pit image is a site on a biochip for binding nucleic acid molecules and generating fluorescent molecules; the first direction and the second direction are perpendicular;
s5: selecting M sub-areas closest to the center of the fluorescent image, calculating the average value of stretching vectors of the M sub-areas in the first direction and the second direction, and calculating the distance of the average value of the stretching vectors, wherein the distance represents the point roundness of the fluorescent image;
s6: calculating the difference between the stretching vector of each sub-area in the first direction and the second direction and the average value of the stretching vectors of the M sub-areas in the first direction and the second direction to obtain a vector difference value, wherein the maximum length value of the vector difference value in the N sub-areas is the definition of the fluorescent image;
s7: comparing the point circularity obtained in the step S5 with a preset point circularity threshold value, and if the point circularity is smaller than the preset threshold value, judging that the image is not dithered, wherein the image is used for subsequent analysis; and (3) comparing the definition obtained in the step (S6) with a preset definition threshold, and if the definition is smaller than the preset threshold, judging that the image is clear, wherein the image is used for subsequent analysis.
In a preferred embodiment, the biochip is a gene sequencing chip and the fluorescent image is a fluorescent image taken of the sequencing chip during nucleic acid sequencing. The biochip can be continuously photographed by a microscope camera to obtain a plurality of fluorescence images, and the fluorescence images can also be obtained by a scanning technology. In this context, "fluorescence image" refers to an image of one field of view taken by a microscopic camera at a time. The microscopic camera has a smaller field of view, and can capture multiple fluorescence images for the biochip during the sequencing process. Each field of view is filled with sites, preferably evenly distributed, for binding nucleic acid molecules that incorporate the nucleotide molecules and generate fluorophores during the sequencing reaction.
According to a preferred embodiment, the pretreatment comprises: and extracting the intensity of each pixel point in the image, and performing normalization operation, wherein the normalization operation comprises but is not limited to methods such as linear function normalization, nonlinear function normalization, L2-norm normalization, zero mean normalization and the like. For example, normalization may also be performed by subtracting the mean value divided by the variance, which removes the mean luminance value (intensity) of the image, i.e., calculates the image mean and image variance for the entire sequenced image, and then subtracts the image mean first for each pixel of the sequenced image, and then divides by the image variance.
In a preferred embodiment, the normalized image is divided and divided into N sub-regions, where N is greater than or equal to 2, and for example, N may take on an even number of values 2,4,6,8, 10, 12, 14, 16, …, etc. For each sub-region, discrete fourier transform is performed on the sub-region, and the discrete fourier transform mainly comprises: and acquiring a frequency domain image of the normalized image, wherein the index of the intensity of gray level change in the frequency domain image is an index of gray level on a plane space. Acquiring the frequency domain image, and obtaining a local maximum value set of a high-frequency region in the frequency domain image; the local maximum set of the high frequency region appears as a local maximum of the high frequency region in the frequency domain image, and the maximum point is centrosymmetric with respect to the frequency domain image. Fig. 5 shows a set of high-frequency local maxima in a frequency domain image, wherein there are 4 main high-frequency local maxima at the top left corner, top right corner, top edge, left edge, respectively, each of which has a high-frequency local maxima point that is centrosymmetric with respect to a zero frequency point at the center of the image.
In a preferred embodiment, for each sub-region, from among the above-mentioned several high-frequency local maxima, 4 non-centrosymmetric high-frequency local maxima are selected, which are converted into stretching vectors of the corresponding pit image in the first direction and the second direction in the pixel coordinate system, and mainly include: under the pixel coordinate system, two high-frequency local maxima representing the included angles of the first direction and the second direction are subtracted to obtain an inclined stretching vector in the included angle direction of the first direction and the second direction, wherein the included angle between the inclined stretching vector and the first direction is 0-90 degrees, preferably 30-60 degrees, more preferably 45-50 degrees. Under a pixel coordinate system, two high-frequency local maxima representing the first direction or the second direction are subtracted to obtain a horizontal vertical vector of the first direction or the second direction, and an included angle between the horizontal vertical vector and the first direction is 0 degree or 90 degrees.The first direction and the second direction are perpendicular to each other. Taking fig. 5 as an example for specific illustration, a frequency doubling point of the upper left corner and the upper right corner of the zero frequency point is related to a vector of the pit image in the direction of an included angle between the first direction and the second direction, and the included angle vector is a vector of the pit image in the direction of an included angle of 45 degrees; the doubling points to the left and above the zero frequency point are related to the vectors of the pit image in the first direction or the second direction. Subtracting a frequency doubling point of the upper left corner of the zero frequency point from a frequency doubling point of the upper right corner of the zero frequency point to obtain a difference value to represent an oblique stretching vector; subtracting a frequency doubling point on the zero frequency point from a frequency doubling point on the left side of the zero frequency point to obtain a difference value, and representing a horizontal and vertical stretching vector. And correspondingly assigning a value which is greater than or equal to 0 in the horizontal and vertical stretching vectors to the vectors of the pit image in the first direction, and correspondingly assigning a value which is less than 0 in the horizontal and vertical stretching vectors to the vectors of the pit image in the second direction. Vector of pit image in first direction, added with value of corresponding position in oblique stretching vector
Doubling; in the vector of the pit image in the second direction, the value of the oblique stretching vector is larger than or equal to 0, and the +.>
Doubling; in the vector of the pit image in the second direction, the diagonal stretching vector value is smaller than 0, and the corresponding position in the diagonal stretching vector is correspondingly subtracted +.>
Multiple times.
According to a preferred embodiment, M sub-regions closest to the center of the fluorescence image are selected, the vector average of the M sub-regions in the first direction and the second direction is calculated, and the length of the vector average is calculated, wherein the length of the vector average can represent the dot circularity of the fluorescence image, and M is smaller than or equal to N. For example, when the image is divided into 8 x 8 sub-regions, M may take a value of 4, and after calculating the vector average of the 4 sub-regions closest to the center of the initial fluorescence image in the first direction and the second direction, the distance of this vector average is calculated, including, but not limited to, euclidean distance, manhattan distance, chebyshev distance, min Shi distance, normalized euclidean distance, cosine similarity, mahalanobis distance, correlation distance, jacobis distance, and the like.
According to a preferred embodiment, comparing the obtained point roundness with a preset point roundness threshold, and if the obtained point roundness is smaller than the preset threshold, judging that the image is not dithered, wherein the image is used for subsequent analysis; if the average value of the vectors obtained in the step S5 is larger than the threshold value, drawing a vector diagram, observing the change of the vector diagram, and correcting the photographing flow of the instrument according to the change.
According to a preferred embodiment, the difference between the vector of each sub-region in the first direction and the second direction and the vector average value of the M sub-regions in the first direction and the second direction is calculated to obtain a vector difference value, and the length maximum value of the vector difference value in the N sub-regions is used for representing the definition of the image. Comparing the obtained definition with a preset definition threshold, and judging that the image is clear if the obtained definition is smaller than the corresponding threshold, wherein the image is used for subsequent analysis; if the image is larger than the definition threshold, drawing a vector diagram according to the obtained N vector difference values, observing the change of the vector diagram, and correcting the photographing flow of the instrument according to the change.
A second aspect of the present invention provides a non-transitory computer readable storage medium having stored thereon a computer program, characterized in that the computer program, when executed by a processor, implements a method of evaluating fluorescence image sharpness and/or dot circularity as any of the methods described herein.
Each of the features discussed in the detailed description of the first aspect of the invention are equally applicable to the detailed description of the non-volatile computer-readable storage medium of the invention. As indicated above, all other features are not repeated here and should be considered to be repeated by reference. Those of ordinary skill in the art will understand how features identified in these implementations can be readily combined with basic feature sets identified in other implementations.
A third aspect of the invention provides a gene sequencer, characterized in that it comprises a processor for implementing a method of assessing fluorescence image sharpness and/or dot circularity as described herein when executing a computer program stored in a memory.
Example 1
1. Carrying out a sequencing reaction on a gene sequence to be detected by using a gene sequencing chip to obtain a sequencing fluorescent image, wherein the size of the image is width height; the unit is pixel, which is the image width multiplied by the image height;
2. for any sequencing image, firstly performing image normalization operation; the specific operation steps are as follows: calculating an image mean value and an image variance of the whole sequencing image, subtracting the image mean value from each pixel of the sequencing image, and dividing the image mean value by the image variance; FIG. 3 (a) is an original drawing, FIG. 3 (b) is a histogram of the original drawing, FIG. 4 (a)
Is a normalized image, and fig. 4 (b) is a histogram of the normalized image.
3. Dividing the normalized image into N sub-areas, wherein N is more than or equal to 2; in this example, the normalized image is equally divided into 64 sub-regions, where N is 64. Then, performing discrete Fourier transform on the image of each sub-region, and moving a zero frequency point to the middle of the frequency spectrum, which is useful for observing the Fourier transform, and acquiring a high-frequency local maximum (a frequency doubling point) in the frequency domain image; as shown in fig. 5.
4. Correspondingly converting the high-frequency local maximum value into a stretching vector of the pit image in a first direction and a second direction under a pixel coordinate system, wherein the first direction and the second direction are perpendicular. In this example, a frequency doubling point of the upper left corner and the upper right corner of the zero frequency point is related to a vector of the pit image in the direction of an included angle between the first direction and the second direction, and the included angle vector is a vector with the pit image in the direction of an included angle of 45 degrees; a frequency doubling point on the left side and the upper side of the zero frequency point is related to the vector of the pit image in the first direction or the second direction; as shown in fig. 5. In this example, subtracting one frequency multiplication point of the upper right corner of the zero frequency point from one frequency multiplication point of the upper left corner of the zero frequency point to obtain a difference value to represent an oblique stretching vector, wherein the vector size is N rows and 1 columns or 1 row and N columns, and the vector type is floating point type(integer); subtracting one frequency multiplication point on the left side of the zero frequency point from one frequency multiplication point on the upper side of the zero frequency point to obtain a difference value, wherein the difference value represents a horizontal and vertical stretching vector, the vector size is N rows and 1 columns or 1 row and N columns, and the vector type is floating point type (integer). Corresponding to a value greater than or equal to 0 in the horizontal and vertical stretching vectors, and assigning the value to the vector of the pit image in the first direction; assigning a value smaller than 0 in the horizontal and vertical stretching vectors to the vectors of the pit image in the second direction correspondingly; vector of pit image in first direction, added with value of corresponding position in oblique stretching vector
Doubling; in the vector of the pit image in the second direction, the value of the oblique stretching vector is larger than or equal to 0, and the +.>
Doubling; in the vector of the pit image in the second direction, the diagonal stretching vector value is smaller than 0, and the corresponding position in the diagonal stretching vector is correspondingly subtracted +.>
Multiple times. Table 1 shows the values of 64 sub-areas in a first direction, here converted into a matrix representation, and table 2 shows the values of 64 sub-areas in a second direction, here converted into a matrix representation.
TABLE 1.64 values of sub-regions in the first direction
TABLE 2 values of 64 sub-regions in the second direction
5. The magnitude of the stretching component of the pit image in the first direction and the second direction determines the degree of ellipse of the pit image, M sub-areas closest to the center of the fluorescent image are selected, M is smaller than or equal to N, the vector average value of the M sub-areas in the first direction and the second direction is calculated, and the distance is calculated, wherein the distance represents the point roundness of the fluorescent image. In the example, 4 sub-areas closest to the fluorescence center are selected, namely M is equal to 4, and vector average values of the 4 sub-areas in the first direction and the second direction of the pit image are calculated; in this example, the distance method for calculating the vector average value is calculated by using the euclidean distance method, and the distance represents the roundness of the image point. Table 3 shows the vector average of M sub-regions, which is converted into matrix representation, the length of the calculated vector average is 0.3349, and the length is compared with a preset threshold value of 0.4 and is smaller than the threshold value of 0.4, and the image meets the requirements and can be normally processed and analyzed.
TABLE 3 vector average of 4 sub-regions
| First direction | Second direction |
| -0.1476 | 0.3006 |
6. And subtracting vector average values of the M sub-areas in the first direction and the second direction correspondingly from vectors of each sub-area in the first direction and the second direction, and obtaining the distance of each sub-area, wherein the maximum value of the distances in the N sub-areas is used for representing the focusing definition of the image. In this example, the distance calculation method adopts a euclidean distance calculation method. Table 4 shows the vector lengths of N sub-regions, here converted to a matrix representation, where the maximum is 1.2840, the preset threshold is 1.2, and the maximum is greater than the threshold; at this time, a vector diagram is drawn according to the mean value vector and the difference value vector in the calculation process, as shown in fig. 6. In the vector diagram, the length of the arrow indicates the vector length, i.e., the magnitude of the focus value, and the direction of the arrow indicates the direction of the vector. From the graph, the lower right corner of the image has larger focusing value and the upper right corner and the lower left corner have smaller focusing value, so that the chip position can be adjusted and the lower right corner chip position can be slightly lifted. The vector diagram is used for judging the shaking direction of the roundness difference of the image points and the degree of virtual focus, and correcting the photographing scheme of the camera.
TABLE 4 vector length for 64 sub-regions
The above is only a preferred embodiment of the present invention, and those skilled in the art can also make alterations and modifications to the above described embodiment, therefore, the present invention is not limited to the above described embodiment, and any obvious improvements, substitutions or modifications made by those skilled in the art on the basis of the present invention are all within the scope of the present invention.
Claims (9)
1. A method of evaluating the sharpness and roundness of a fluorescent image, comprising the steps of:
s1, acquiring a fluorescence image shot by the biochip at a preset shooting position;
s2, preprocessing the fluorescent image to obtain a preprocessed image;
s3, dividing the preprocessed image into N sub-areas, wherein N is more than or equal to 2; performing discrete Fourier transform on each sub-region to obtain a local maximum value set of the frequency domain image in a high frequency region;
s4, selecting non-centrosymmetric 4 high-frequency local maxima from the local maxima set, and converting the non-centrosymmetric 4 high-frequency local maxima into stretching vectors of corresponding pit images in a first direction and a second direction under a pixel coordinate system; the pit image is a site on a biochip for binding nucleic acid molecules and generating fluorescent molecules; the first direction and the second direction are perpendicular;
s5: selecting M sub-areas closest to the center of the fluorescent image, calculating the average value of stretching vectors of the M sub-areas in the first direction and the second direction, and calculating the distance of the average value of the stretching vectors, wherein the distance represents the point roundness of the fluorescent image;
s6: calculating the difference between the stretching vector of each sub-area in the first direction and the second direction and the average value of the stretching vectors of the M sub-areas in the first direction and the second direction to obtain a vector difference value, wherein the maximum length value of the vector difference value in the N sub-areas is the definition of the fluorescent image;
s7: comparing the point circularity obtained in the step S5 with a preset point circularity threshold value, and if the point circularity is smaller than the preset threshold value, judging that the image is not dithered, wherein the image is used for subsequent analysis; and (3) comparing the definition obtained in the step (S6) with a preset definition threshold, and if the definition is smaller than the preset threshold, judging that the image is clear, wherein the image is used for subsequent analysis.
2. The method of claim 1, wherein the surface of the biochip has a plurality of micropits, which are sites where fluorescence occurs, and the period of the micropits remains uniform after imaging; the pit period is the number of pixels that a single pit occupies after imaging.
3. The method of claim 1, wherein the preprocessing in S2 includes normalization operations including, but not limited to, linear function normalization, nonlinear function normalization, L2-norm normalization, zero-mean normalization, and the like.
4. The method of claim 1, wherein the converting the high frequency local maxima into the stretching vectors in the first direction and the second direction in the pixel coordinate system in S4 comprises: under a pixel coordinate system, subtracting two high-frequency local maxima representing included angles of a first direction and a second direction to obtain an oblique stretching vector in the included angle direction of the first direction and the second direction, wherein the included angle between the oblique stretching vector and the first direction is 0-90 degrees; under a pixel coordinate system, two high-frequency local maxima representing the first direction or the second direction are subtracted to obtain a horizontal vertical vector of the first direction or the second direction, and an included angle between the horizontal vertical vector and the first direction is 0 degree or 90 degrees.
5. The method of claim 4, wherein the oblique stretching vector is preferably at an angle of 30 ° to 60 °, more preferably 45 ° to 50 °, to the first direction.
6. The method of claim 1, wherein the distances in S5 include, but are not limited to, euclidean distance, manhattan distance, chebyshev distance, min Shi distance, normalized euclidean distance, cosine similarity, mahalanobis distance, correlation distance, jekade distance, and the like.
7. The method as recited in claim 1, further comprising:
s8: comparing the point circularity obtained in the step S5 with a preset point circularity threshold value, if the point circularity is larger than the threshold value, drawing a vector diagram according to the stretching vector average value obtained in the step S5, observing the change of the vector diagram, and correcting the photographing flow of the instrument according to the change; comparing the definition obtained in the step S6 with a preset definition threshold, if the definition is larger than the definition threshold, drawing a vector diagram according to N vector difference values obtained in the step S6, observing the change of the vector diagram, and correcting the photographing flow of the instrument according to the change.
8. A gene sequencer, characterized in that it comprises a processor for implementing the method of evaluating fluorescence image sharpness and dot roundness according to any of claims 1-7 when executing a computer program stored in a memory.
9. A non-transitory computer readable storage medium having stored thereon a computer program, which when executed by a processor, implements the method of assessing fluorescence image sharpness and dot roundness according to any of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310295465.3A CN116309491A (en) | 2023-03-24 | 2023-03-24 | Method for evaluating definition and point roundness of fluorescent image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310295465.3A CN116309491A (en) | 2023-03-24 | 2023-03-24 | Method for evaluating definition and point roundness of fluorescent image |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116309491A true CN116309491A (en) | 2023-06-23 |
Family
ID=86823856
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310295465.3A Pending CN116309491A (en) | 2023-03-24 | 2023-03-24 | Method for evaluating definition and point roundness of fluorescent image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116309491A (en) |
-
2023
- 2023-03-24 CN CN202310295465.3A patent/CN116309491A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11100637B2 (en) | System and method for calculating focus variation for a digital microscope | |
| US10768403B2 (en) | Auto-focus methods and systems for multi-spectral imaging | |
| US8508588B2 (en) | Methods and systems for identifying well wall boundaries of microplates | |
| Brändle et al. | Robust DNA microarray image analysis | |
| CN113109368B (en) | Glass crack detection method, device, equipment and medium | |
| US6876776B2 (en) | System and method for auto-focusing an image | |
| KR102195826B1 (en) | Keypoint identification | |
| JP2013526717A5 (en) | ||
| US8983199B2 (en) | Apparatus and method for generating image feature data | |
| CN114549600A (en) | Fluorescent image registration method | |
| CN111971711A (en) | Fluorescent image registration method, gene sequencer system and storage medium | |
| US7551762B2 (en) | Method and system for automatic vision inspection and classification of microarray slides | |
| CN110111842B (en) | Image definition analysis and focusing method, sequencer, system and storage medium | |
| CN117252786A (en) | Gene detection data enhancement processing method | |
| CN116309491A (en) | Method for evaluating definition and point roundness of fluorescent image | |
| CN112345548A (en) | Method and device for detecting surface forming smoothness of graphite plate of fuel cell | |
| Pollatou | An automated method for removal of striping artifacts in fluorescent whole-slide microscopy | |
| CN113670865B (en) | Resolution board, resolution evaluation method and related equipment | |
| US10733707B2 (en) | Method for determining the positions of a plurality of objects in a digital image | |
| Bieniecki et al. | Multi-pass approach to adaptive thresholding based image segmentation | |
| CN109470164A (en) | A kind of tracking of view-based access control model has label extensometer dynamic measurement and system | |
| Fouad et al. | Automatic Segmentation of cDNA Microarray Images using different methods | |
| Bradley et al. | Image Processing | |
| KR20220160313A (en) | Method and apparatus for detecting data matrix | |
| RU2177602C1 (en) | Process displaying localization zones of deformation of surface |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |