Development of Abnormal Red Blood Cells Classifier Using Image Processing Techniques with Support Vector Machine

Carlos C. Hortinela, Janette C. Fausto, Flordeliza L. Valiente, Paul Daniel C. Divina, John Philip T. Felices

2020 · 11 citations · 8 references

Concepts

Abstract

In some studies conducted from the medical field, the abnormality in the morphological structure of red blood cells is used as an indicator of blood related diseases and conditions based on the classification of the abnormal red blood cells that have been found. The system provides an automated way of identification and classification of abnormal red blood cell based on morphology. This is done by capturing the microscopic image of a blood smear and applying image processing techniques. The captured image undergoes several processes such as image pre-processing, segmentation, and feature extraction. The system acquires important morphological properties of red blood cells through these processes. Properties such as diameter, shape geometric factor, central pallor and target flag are the main parameters used by the system. Support vector machine is applied as classifier to identify the type of the abnormal cell. The SVM will be able to do the classification by training it with dataset for each class of red blood cell included in the system. These datasets are images of the blood cells that have parameters within the range of a specific class. The system also provides a list of blood related diseases and condition associated with the abnormal red blood cells found to be present. Based on the result of the testing, the system was able to produce a high accuracy output of 96.67% for normal RBC, 100% for echinocytes, 97.5% for elliptocytes, 97.5% for dacrocytes, 98.33% for spherocytes, 97.5% for Target cells, 98.53% for stomatocytes and 86.67% for unknown class.

References

8