<![CDATA[The nuclei and cytoplasm can be detected through Pap smear images. The image consists of cytoplasm and nuclei. In Pap smear image, nuclei are the most critical cell components and undergo significant changes in cervical cancer disorders. To help women avoid cervical cancer, early detection of nuclei abnormalities can be done in various ways, one of which is by separating the nuclei from the non-nucleis part by image segmentation it. In this study, segmentation of the separation of nuclei with other parts of the Pap smear image is carried out by applying the U-Net CNN architecture. The amount of pap smear image data is limited. The limiter data can cause overfitting on U-Net CNN model. Meanwhile, U-Net CNN needs a large amount of training data to get great performance results for classification. One technique to increase data is augmentation. Simple techniques for augmentation are flip and rotation. The result of the application of U-Net CNN architecture and augmentation is a binary image consisting of two parts, namely the background and the nuclei. Performance evaluation of combination U-Net CNN and augmentation technique is accuracy, sensitivity, specificity, and F1-score. The results performance of the method for accuracy, sensitivity, and F1-score values are greater than 90%, while the specificity is still below 80%. From these performance results, it shows that the U-Net CNN combine augmentation technique is excellent to detect nuclei in compared to detect non nuclei cell on pap smear image.]]>
