Morphological Feature Extraction from Low Quality of Thick Blood Smear Microphotographs

Identification of the parasite presence in a blood smear is the most crucial part of the diagnosis of malaria. In the case of the low quality of smear, the diagnosis process becomes more complex because the parasites are difficult to read. The identification accuracy can be improved using morphological features that microscopist commonly used in diagnosing malaria. In this paper, we proposed a strategy to reduce false-positive identification of malaria parasites using rule-based on the features. This paper shows how to extract the important features and determine the rule of identification of malaria parasites. Moreover, a splitting algorithm was developed for cases where core and cytoplasm are joined in segmentation result. The proposed method was tested on 50 candidate parasite. The experimental results show that the proposed method can reduce false positives greater than 70% compared to the results of the previous system.