Effectiveness of Convolutional and Capsule network in Malaria Parasite Detection
Ali Sayyed, Dipayan Saha, Abdul Rakib Hossain, Celia Shahnaz · 2019
Malaria, a life-threatening disease, takes millions of lives every year all over the world. Early detection of malaria can be an effective way to right with this disease. This study investigates the effectiveness of the combination of convolutional and capsule network to detect the malaria parasite in red blood cells in terms of standard evaluation metrics. The proposed architecture achieves higher accuracy in comparison to different existing convolutional neural networks. The proposed method also shows stable and higher accuracy in case of different ratios of train and test set separation. The proposed architecture attains about 95% accuracy when only 10% of the dataset is kept for training purpose. Such performance ensures the superiority of the CNN-Capsule network in malaria parasite detection.