Analysis of Convolutional Neural Networks and Shape Features for Detection and Identification of Malaria Parasites on Thin Blood Smears

The gold standard for malaria diagnosis still remains to be microscopy. However, cases from remote areas needing immediate diagnosis and treatment can benefit from a faster diagnostic process. Several intelligent systems for malaria diagnosis have been proposed using different computer vision techniques. In this research, models using convolutional neural networks, and a model using extracted shape features are implemented and compared. The CNN models, one trained from scratch and the other utilizing transfer learning, with accuracies of 92.4% and 93.60%, both outperform the shape feature model in malaria parasite recognition.

[1]  Dumitru Erhan,et al.  Going deeper with convolutions , 2014, 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[2]  Prospero C. Naval,et al.  Malaria Parasite Detection and Species Identification on Thin Blood Smears Using a Convolutional Neural Network , 2017, 2017 IEEE/ACM International Conference on Connected Health: Applications, Systems and Engineering Technologies (CHASE).

[3]  Chetan Gupta,et al.  Detection of malaria parasite species and life cycle stages using microscopic images of thin blood smear , 2016, 2016 International Conference on Inventive Computation Technologies (ICICT).

[4]  T. Limpiti,et al.  Automated classification of malaria parasite species on thick blood film using support vector machine , 2015, 2015 8th Biomedical Engineering International Conference (BMEiCON).

[5]  Raghuveer M. Rao,et al.  Segmentation of malaria parasites in peripheral blood smear images , 2009, 2009 IEEE International Conference on Acoustics, Speech and Signal Processing.