Mosquito Classification Using Convolutional Neural Network with Data Augmentation

Mosquitoes are responsible for the most number of deaths every year throughout world. Bangladesh is also a big sufferer this problem. Dengue, malaria, chikungunya, zika, yellow fever etc. caused by dangerous mosquito bites. The main three types mosquitoes which found in aedes, anopheles and culex. Their identification crucial to take necessary steps kill them an area. Hence, convolutional neural network (CNN) model developed so that could be classified from their images. We prepared local dataset consisting 442 images, collected various sources. An accuracy 70% has been achieved running proposed CNN on dataset. However, after augmentation becomes 3,600 increases 93%. showed comparison some methods with method VGG-16, Random Forest, XGboost SVM. Our outperforms these terms classification mosquitoes. Thus, research forms example humanitarian technology, where data science can used support classification, enabling treatment borne diseases.