Robust Illumination Compensation Method For Microscope Images

In recent years,cancer has appeared more and more frequently,showing an outbreak trend.The 2018 National cancer report showed that 7 people per minute in our country are diagnosed with cancer and 5 died of cancer.Early detection of early treatment is the most effective way to treat cancer,but manual examination alone cannot meet this requirement.Image processing and analysis is a main tool for DNA ploidy analysis.But microscope images are often not ideal;images specifically often suffer from an uneven background due to illumination flaws.Consequently,for proper images analysis,background correction methods are required.In this paper we proposed two methods based on background image.Both methods estimate illumination through images without cells,one method uses interpolated function,the other uses more than one image under the same illumination condition to filling the background.Then a method called mean deviation estimate is used for both methods to estimate the illumination bias,an image called bias image is obtained through which.At last the uneven illumination image and the bias image are added to get the final even illumination image.According to the experiments,the proposed methods performed better than the existed popular methods provided by CellProfiler.Accurate measurement of the relative content of DNA is a prerequisite for accurate diagnosis of ploidy analysis system.However,no matter using halogen lamp or LED lamp,the problem of uneven illumination is unavoidable.The optical measurement depends on the gray value,the concentration of DNA in the different positions of the same measurement cell results are not the same,or even far.This will directly increase the measurement error of the system and reduce the credibility of the ploidy measurement system's results.Therefore,it is necessary to reduce or eliminate the effect of uneven illumination through effective illumination compensation.Traditional illumination compensation methods depend on clean background glass without cells.Each time,additional glass is used to compensate illumination,which leads to low robustness.In addition,traditional methods cannot cope with changing illumination conditions,which has a great impact on the efficiency of illumination compensation.In order to improve the accuracy of DNA ploidy analysis technology,this paper improves the quality of cell image from three aspects: background filling,bias estimation and light compensation.Specific work includes the following points:1.In order to solve the problem of inaccurate estimation of illumination deviation,an effective deviation estimation method is proposed.In this method,mean deviation estimation and Bayesian minimum mean square error estimation are proposed for background estimation.This method only uses the glass to compensate illumination,which improves the current situation of using additional glass to estimate background bias.It is characterized by using the illumination characteristics of the image to be compensated to accurately estimate the gray value of each position of the deviated image.Compared with the existing methods,this method is more convenient to operate,and the background estimated by this method and the subsequent IOD measurement results are more accurate.2.In order to solve the situation that it is impossible to distinguish impurities and cells in the background acquisition link,an illumination compensation method based on interpolation filling and cumulative compensation is proposed.Firstly,the interpolation method is used to fill the cell part with the same gray value as the surrounding background.Then,the accumulated image is used in the scanning process to find the gray value at the same position as the compensated cell to fill in the compensated position.This method adapts to the change of illumination and can distinguish cells and impurities more effectively.Compared with traditional methods,it has faster speed and better compensation effect.After that,the IOD measurement effect is more accurate.3.In order to solve the change of illumination with time in the scanning process,an illumination compensation method based on forgetting factor is proposed.On the basis of the above cumulative compensation,the bias background is estimated by adding weights using multiple cumulative images.This method maximizes the use value of the scanned image,makes it more suitable for changing illumination conditions,makes the final compensation effect more accurate and more in line with human vision,and further improves the IOD measurement results.Experiments show that the robust illumination compensation method proposed in this paper is more accurate in background estimation and greatly improves the processing speed compared with the existing illumination compensation methods.In this paper,interpolation filling and accumulative compensation method are proposed,which can effectively improve the accuracy of measurement under the condition that the image clarity is guaranteed to the greatest extent.The proposed illumination compensation method with forgetting factor can effectively improve the uniformity of image gray value and the accuracy of final measurement.