Automatic Colour Staining on Slide Histopathology Images using RGB Image Processing

Abstract

Nowadays, staining on slide histopathology images is widely utilized in medicine. The histopathology analysis is conducted manually by the pathologist, including tissue structure, distribution of cells in tissues and others. However, since this assessment is generally performed visually by pathologists, it can suffer from significant inter-observer variability. To overcome this difficulty, automatic staining on slide histopathology using assisted image processing provided a more quantitative diagnosis of cells. This is a step in preparing automatic staining on slide histopathology images using image processing. Firstly, convert the original picture from the hospital to a grayscale picture since the original picture from the hospital is overly bright using the ImageJ application. ImageJ converts the original picture from the hospital to grayscale by setting the image to 32 bits and adjusting the brightness and contrast. Then, after obtaining the grayscale picture, the second step is converting the grayscale to an RGB colour. This step is implemented using the Matrix Laboratory (MATLAB) coding application and then the comparison result of the grayscale picture and the final picture from manual staining is obtained. For the result analysis, this research used an online application to examine the colour similarity between final manual staining and automatic staining using MATLAB. The similarity of colour staining by MATLAB exceeds 50% for all samples compared to manual staining. In conclusion, the application of automatic staining on slide histopathology using an image processing technique is a much better method to classify each cell.