CN106558041A - The band probe algorithm suppressed based on local in gel electrophoresiss digital picture - Google Patents

Abstract

The invention discloses a method for automatically extracting swimming lane features in gel electrophoresis digital images. First construct the characteristic curve of the brightness change in the lane column direction, then calculate the average height and variance of the curve, calculate the initial maximum value according to the definition of the extreme point, the relative height of the extreme point and the distance between the extreme points, and use the average height and variance Filter these maximum points, retain the points as the initial bright bar position, and then divide the characteristic curve into several segments according to these positions, suppress each segment according to the average height and variance of each segment, and then combine them to construct a new The characteristic curve, and finally the maximum values are calculated on the new characteristic curve, and these maximum values are used as the final strip positions. The invention can not only obtain the position of the real strip, but also filter out the false strip, and can also improve the recognition rate of the low-brightness strip, greatly reduce the times of manual intervention, and improve the working efficiency.

Description

Band Detection Algorithm Based on Local Suppression in Digital Image of Gel Electrophoresis

technical field

The invention belongs to the technical field of image processing, and relates to a method for automatically detecting bands from swimming lane graphics by using region suppression technology to process gel electrophoresis digital images.

Background technique

No relevant literature has been found on the processing of gel electrophoresis images, but some relevant literature appears in other fields. Chao Yang, Zengyou He and Weichuan Wu[Chao Yang, ZengyouHe, WeichuanWu.Comparison of public ic peak detection algorithms for MALDI massspectrometry data analysis, BMC Bioinformatics2009, 10:4] extracted signal peaks through three parts, first smoothing the signal, and then Calculate the baseline for correction, and finally perform peak detection. The peak detection method uses a variety of methods, including threshold method, peak slope method, local extremum method, ridge line method, etc. Hsein-Ping Kew with Do-Un Jeong [Hsein-Ping Kew, Do-Un Jeong. Variable Threshold Method for ECG R-peak Detection, J MedSyst (2011) 35: 1085-1094] using differential technique and Hilbert transform pair The ECG curve is preprocessed, and then a multi-threshold method is constructed to extract the peak value of the signal. S. Satheeskumaran and M. Sabrigiriraj [S. Satheeskumaran, M. Sabrigiriraj. A New LMS Based Noise Removal and DWT Based R-peak Detection in ECG Signal for Biotelemetry Applications. Natl. Acad. Sci. Lett., 2014, 37(4): 341 -349.] When processing the ECG curve, the curve is first filtered with an adaptive filter based on least squares, and then the peak value of the ECG is detected using wavelet analysis.

Contents of the invention

The purpose of the present invention is to solve the problem of automatic detection of characteristic bands in the image. For this reason, the present invention proposes a method based on region suppression to automatically detect the position information of the swimming lane bands, and apply it to the digital image processing of gel electrophoresis . This method fully considers the characteristics of gel electrophoresis digital images, that is, there are regions, and the regional characteristics are very different, so a separate processing strategy is adopted, and good results are achieved.

In order to complete the automatic detection technology of swim lane strips, a technical solution based on region suppression is proposed, as follows:

Step 1: Construct the characteristic curve, and detect as much as possible, and divide the region for the characteristic curve;

Step 2, calculate the regional characteristics of the curve, and perform local suppression;

Step three, recombining the suppressed characteristic curves to construct new characteristic curves for precise detection.

The implementation process is described in detail below:

Step S1, constructing a characteristic curve for the swimlane graph. The construction method is to count the average gray value of each row in the swimlane image as the y value, and use the position in the column direction as the x value, and combine the two to construct a characteristic curve.

Step S2, using the discrete function extreme point calculation method to detect the extreme point position of the characteristic curve.

Step S3, using different filtering conditions to filter the initial extreme points to obtain the initial strip position, for example, according to the overall statistical characteristics (expectation and variance) of the curve, remove the extreme points below a certain height, according to the adjacent The distance of the extreme point, remove the extreme point with smaller function value.

Step S4, divide the characteristic curve of the lane into regions according to the initial strip position.

In step S5, the curve characteristics of each region are counted, and the expectation and variance thereof are calculated.

Step S6, suppress the region according to the expectation and variance, that is, subtract a certain height value from all function values.

Step S7, integrating the suppressed segmented curves to obtain a new characteristic curve.

In step S8, steps S2 and S3 are executed to obtain accurate strip positions.

Description of drawings

Figure 1 Algorithm Flowchart: Briefly describes the logic flowchart of the algorithm.

Figure 2 Algorithm Processing Result Diagram: It shows the processing result of the algorithm.

detailed description

Step 1, the lane image is Characteristic curve f(x,y)={(M _i ,i)}, i∈{1,2,...,n}

Step 2, the calculation method of the extreme point of the discrete curve, where g(i,j)=f'(i,j)

Step 3, the expectation and variance of the curve,

Filter condition 1: f<E+αD, α∈[-3, 3], in the calculation process, α=0.3 is better, it is recommended to set it as the default value, and allow users to modify it.

Filter condition 2: the distance between two adjacent points d<5;

Step 4, according to the maximum value point set obtained in step 3, divide the interval of the function into k sub-regions, which are [1, n ₁ ], [n ₁ +1, n ₂ ], ..., [n _{k -1} +1, n].

Step 5, according to the formula in step 3, calculate the expectation and variance of each segment of the function respectively, denoted as E _k , D _k ;

Step 6, the maximum suppression method of the first region, h _l (i, j) = g _l (i, j)-(E _k +βD _k ), β∈[-3,3], and then k h _l (i, j) are combined to form a new characteristic curve h(i, j). In the experiment, β=1 has a better effect. It is recommended to set it as the default value and allow users to modify it.

Step seven, calculate the position of the strip on h(i, j) according to steps two and three.

The above is only the specific implementation method of the method, but the protection scope of the invention is not limited thereto. Anyone who is familiar with and understands the technology can understand or think of the transformation of some parameters within the scope disclosed by the present invention. For example, α, β, and the distance between two maximum values, etc., should all be covered within the scope of the present invention. Therefore, the protection scope of the present invention should be determined by the protection scope of the claims.

Claims (5)

1. The method for automatically detecting the bands in the swimming lane by using the method of regional suppression, its characteristics include the following aspects: First, the introduction of secondary detection technology, a rough detection, a fine detection; Second, the regional suppression technology is introduced to highlight the true peak and suppress the false peak; Third, different regions use different indicators for suppression. 2. The automatic detection method according to claim 1, the first rough detection determines the position of the main bands, and divides the characteristic curve of the swimming lane into different regions according to these positions. 3. According to claim 1 and claim 2, the statistics of different regions are calculated and used as the parameters of signal suppression in this section. 4. According to claim 1, the area suppression technology makes the peak of the area more prominent, and the low peak is smaller. This technology is more effective for the low-brightness band at the end of the swimming lane, and can distinguish the low-brightness band from the background . 5. The region suppression method as claimed in claim 1 is suitable for automatic detection and processing of swimming lane bands in gel electrophoresis digital images.