International Journal of Computer Science Issues

More than a traditional journal...

This paper presents the research on analysis and extraction of features of red blood cells for anemia recognition. Images from the blood samples collected at a hospital were used. Three geometrical features were used to distinguish between normal and anemic cells; Fourier descriptors, aspect ratio and moments. The City block distance measure was used as a criterion to determine the similarity degree between the tested samples and the established templates. Test results indicate that combination geometrical features gave high discriminative power approaching 98%.