Object detection technique for malaria parasite in thin blood smear images

The infected red blood cell pixel count in thin blood smear image plays a vital role in malaria parasite detection analysis. This paper proposes three stage object detection procedure of computer vision with Kernel-based detection and Kalman filtering process to detect malaria parasite. The use of Kernel based detection with exact pixel information makes the proposed procedure capable of accurately detecting and localizing the target infected by malaria parasites in thin blood smear images. The experiment is conducted on several microscopically preliminary screened benchmark gold standard diagnosis datasets of blood smear images, each 300×300 pixels of Plasmodium falciparum in thin blood smear images. The 300×300 size images were split into overlapping patches, each of size 50×50 pixels. The experimental results on the malaria blood smear image datasets demonstrate the effectiveness of the proposed method over the existing computer vision algorithms. The novelty of the work lies in the application of an object detection for malaria parasite identification in computer vision for thin blood smear images.