Automatic Sickle Cell Anaemia Detection Using Image Processing Technique

Sickle cell anaemia is a congenital red blood cell disorder which depicts an absence of sufficient red blood cells that are fit for transporting oxygen all over body. Usually, ordinary red bloods are disc shaped. But in case of sickle cell anaemia (SCA), the red blood cell is formed in crescent shape instead of disc shapes. A novel methodology is introduced in this study to detect, classify and count the sickle cells available in the red blood cell images to start early identification and treatment. In the developed methodology, the sample images of blood containing both normal and sickle cells are collected. The pre-processing consists of grey scale image conversion and noise filtering using median filter. Watershed segmentation aids in partitioning the images distinctly so that normal cells and sickle cells can be detected. Region properties are calculated to find the number of sickle cells in the given image. Amongst the various region properties, centroid plays an important role for the detection and counting the number of sickle cells available in the image. This automated recognition will especially be useful to the medical experts as a decision support system.