Automatic Breast Cancer Diagnostic System Using Hidden Markov Model and Modified Backpropagation

This item is published by Universitas Islam Negeri Sunan Ampel Surabaya

Lubab, Ahmad and Rini, Dian Candra and Sawiji, Asri (2018) Automatic Breast Cancer Diagnostic System Using Hidden Markov Model and Modified Backpropagation. Other thesis, UIN Sunan Ampel Surabaya.

[img] Text
Breast Cancer Final Report No WM_Lubab.pdf

Download (10MB)

Abstract

A diagnostic system needed to help doctors deal with illness. Breast cancer is a dangerous disease that can affect anyone, either women or men. Identification of can be done using a mammography tool that produces a mammogram image. In this study, The improvement or image of images in image processing is done using adaptive histogram, then followed by the segmentation process using HMM. HMM is segmented by calculating the probability values between pixels based on neighboring properties, then two dimensions HMM applied by using Viterbi training to get good features. After getting a vector feature from the HMM results, modified backpropagation utilized the use of hidden layer nodes that are randomly used. It aims to make the training process faster by optimizing linear errors and non-linear errors. The system produces the best accuracy of 80%.

Statistic

Downloads from over the past year. Other digital versions may also be available to download e.g. from the publisher's website.

Item Type: Thesis (Other)
Creators:
CreatorsEmailNIM
Lubab, Ahmadahmadlubab@uinsby.ac.id2011118102
Rini, Dian Candradiancrini@uinsby.ac.id2024118502
Sawiji, Asrisawiji.asri@uinsby.ac.id2026068701
Contributors:
ContributionNameEmailNIDN
Author., ...
Subjects: Matematika
Teknologi > Teknologi Informasi
Keywords: Breast Cancer; Hidden Markov Model; Viterbi training
Divisions: Karya Ilmiah > Laporan Penelitian
Depositing User: Editor : Ummir Rodliyah------ Information------library.uinsby.ac.id
Date Deposited: 06 May 2020 01:53
Last Modified: 06 May 2020 02:07
URI: http://digilib.uinsby.ac.id/id/eprint/39758

Actions (login required)

View Item View Item