Classification of Prostate Cells Cancer Staging using VGG-19 and GoogLeNet Models
Yessi Jusman, Shafa Cahyaningtyas, Feriandri Utomo · 2023
Prostate cancer attacks the men’s urinary system. Many studies on prostate cancer and other diseases have utilized artificial intelligence (AI). Moreover, evaluating the models employed in disease classification is crucial to serving as a reference for other researchers in determining an appropriate model. This study applied two preserved models: VGG-19 and GoogLeNet. With an accuracy of 95.47% during training, VGG-19 outperformed GoogLeNet, acquiring 94.44%. Nevertheless, compared to VGG-19, GoogLeNet’s training time was significantly shorter at an average of 7 minutes and 46 seconds. In contrast, VGG-19 required a relatively long time, clocking in at 56 minutes and 38 seconds. VGG-19 also performed better in testing than GoogLeNet, unveiled by a confusion matrix assessment. In other words, VGG-19 performed significantly better than GoogLeNet in classifying prostate cell images.