Title :
Classification of three types of red blood cells in peripheral blood smear based on morphology
Author :
Soltanzadeh, Ramin ; Rabbani, Hossein
Author_Institution :
Dept. of Biomed. Eng., Isfahan Univ. of Med. Sci., Isfahan, Iran
Abstract :
Analysis of microscopic images is used in many fields of technology and medicine. In some medical experiments, some drugs, with known effects in red blood cells (R.B. Cs) membranes, are used to find out their activity. In addition, different diseases affect the appearance of R.B. Cs and cause the malfunction of the internal organs. In this paper, Elliptocytes, Discocytes and Echinocytes, which are three known R.B. Cs, are classified in a peripheral blood smear based on morphological methods. For this reason, a simple statistical analysis of distances of each edge pixel from mass center is employed. This method is 98.63% successful for Elliptocyte recognition, 96.7% successful for the normal Discocyte recognition and 95.36% successful for Echinocyte recognition.
Keywords :
blood; image classification; image recognition; mathematical morphology; medical image processing; microscopy; statistical analysis; discocyte recognition; echinocyte recognition; elliptocyte recognition; image classification; microscopic images; morphological methods; peripheral blood smear; red blood cells; statistical analysis; Cells (biology); Feature extraction; Image edge detection; Microscopy; Pixel; Red blood cells; Discocytes; Echinocytes; Elliptocytes; Morphological method; Peripheral blood smear;
Conference_Titel :
Signal Processing (ICSP), 2010 IEEE 10th International Conference on
Conference_Location :
Beijing
Print_ISBN :
978-1-4244-5897-4
DOI :
10.1109/ICOSP.2010.5655754