Mosquito Type Identification using Convolution Neural Network

Mosquitoes are insects which are most commonly found in places with high water content. There are various types of mosquitos and each type has its own destructive characteristics. In order to completely eradicate mosquitoes from a particular place, people must know what type of mosquito it is. But this is not a task that can be done with naked eyes. To help this situation, this study develops a Deep Learning (DL) model which can predict the type of mosquito with great accuracy. For this purpose, a dataset of images of the Aedes and culex species of mosquito is collected from Kaggle. Following that, this dataset is split into three parts: training, validation, and testing. Then, pixel and image scaling is used to preprocess this dataset. A DL model is produced using the Convolutional Neural Network (CNN) technique. The model is then trained using the preprocessed dataset. The results of the training are also calculated and documented. The results are then put into a graph for easier examination. The highest accuracy of the model is produced during training and it is 99.2%. As for the loss values, the lowest loss value is 0.8% which is so small for a classification algorithm. Next, tests are conducted on the trained and validated models. A confusion matrix is then used to examine the model performance. The model is then assessed using the metrics such as accuracy, True Positive Rate (TPR), True Negative Rate (TNR), False Positive Rate (FPR), False Negative Rate (FNR), precision, and F1 score. These parameters’ final values were also determined to be satisfactory.