Leukemia Cell Image Classification Using CNN: AlexNet and GoogLeNet

Leukemia is a type of blood cancer that often causes worry and uncertainty for sufferers. This study aims to investigate the latest developments in the use of convolutional neural network (CNN) models, especially GoogLeNet and AlexNet, in classifying leukemia cell images. In this study, both models were trained and tested using previously trained leukemia image datasets. The results showed that both GoogLeNet and AlexNet had very good performance in classifying leukemia cell images. Both models achieved 100% accuracy in each training run, demonstrating outstanding ability to identify and differentiate the images. In addition, both models also show equally good results in the assessment using the confusion matrix, including precision, sensitivity, specificity, and F-score. In addition to accuracy, a comparison was also made of the time parameters required in model training. In this case, AlexNet is proven to be more efficient with a shorter training time than GoogLeNet. This shows that AlexNet can be a better choice in terms of time efficiency in model training. This research provides hope in dealing with leukemia by using deep learning technology and convolutional neural network models. The results obtained indicate the possibility of using this model in the diagnosis and classification of leukemia images accurately and efficiently. However, further research is needed to validate and test this model on a larger dataset and consider other factors that may affect model performance.