Skripsi

SEGMENTASI NUKLEUS KANKER SERVIKS PADA CITRA PAP SMEAR DENGAN MENGGUNAKAN U-NET CONVOLUTIONAL NEURAL NETWORK (CNN)

---

Penilaian

---

Pap smear image from Zenodo dataset consists of nucleus and cytoplasm. The nucleus is the most important element in the cell and undergoes significant changes in the event of cervical cancer. To prevent cervical cancer in women, early detection of nucleus abnormalities can be done, one of which is by separating the nucleus and non-nucleus in the image. To get the features of the nucleus, a segmentation process is needed. The purpose of this study was to determine the performance results of the CNN U-Net architecture on nucleus segmentation. The measures used for the performance of the model are accuracy, sensitivity, specificity, and F-measure. The Zenodo dataset used is 184 data which can be divided into 98 training data, 25 validation data, and 61 testing data. The method used in this research consists of data collection, architecture implementation, training, testing, and evaluation. The results of the U-Net CNN architecture performance in the nucleus segmentation process are accuracy of 91.39%, the sensitivity of 97.02%, specificity of 73.47%, and F-measure of 94.49%. Based on these results, it can be concluded that the U-Net architecture has succeeded in segmenting the nucleus and predicting non-nucleus (background) pixels very well, this is indicated by the accuracy, sensitivity, and F-measure values above 90%, although the specificity value is between 70 % and 80% are quite good in segmenting pap smear images. Keywords: Images, Segmentation, Cancer Cervical, Nucleus, U-Net

---

Availability

Inventory Code	Barcode	Call Number	Location	Status
2107003457	T58681	T586812021	Central Library (REFERENCES)	Available but not for loan - Not for Loan

Detail Information

Series Title

-

Call Number

T586812021

Publisher

Inderalaya : Prodi Ilmu Matematika, Fakultas Matematika dan Ilmu Pengetahuan Alam., 2021

Collation

x, 59 hlm. : ilus. ; 29 cm

Language

Indonesia

ISBN/ISSN

-

Classification

618.1407

Content Type

Text

Media Type

unmediated

Carrier Type

-

Edition

-

Subject(s)

kanker serviks
Prodi Ilmu Matematika
U-Net Convolutional Neural Network (CNN)

Specific Detail Info

-

Statement of Responsibility

DS

Other version/related

File Attachment

Comments

---

You must be logged in to post a comment