Image preprocessing and hyperparameter optimization on pretrained model MobileNetV2 in white blood cell image classification

White blood cells play a role in maintaining the immune system which consists of several types such as neutrophils, lymphocytes, monocytes, eosinophils and basophils. MobileNetV2 is one pretrained convolutional neural network (CNN) models that provides excellent advantages performance classifying images. In this research was conducted to find out how apply optimization hyperparameters impact image processing on white cell classification using MobileNetV2, so it expected combination preprocessing hyperparameter values can produce highest accuracy value. To maximize process, before image, stages are carried out, namely cropping, grayscale, resizing augmentation. Hyperparameter tuning for an experiment improve model performance. The three main parameters used learning rate, batch size, number epochs. Performance will be measured accuracy, sensitivity, specificity confusion matrix. Based experimental results study, shows best rate value 0.00001, size 32, epoch 250.