Klasifikasi Abnormalitas Sel Darah Merah Untuk Deteksi Dini Myeloproliferative Syndrome Berbasis Neural Network
Total View This Week0
Total View This Week0
Institusion
Politeknik Negeri Jember
Author
Parahita, Syavina Octavia
Subject
458 - Teknik Informatika
Datestamp
2021-10-11 02:28:57
Abstract :
Polycythemia Vera (PV) is one of the categories of Myeloproliferative Neoplasms Syndrome which is characterized by an increase in the number of red blood cells (erythrocytosis) as well as an increase in hemoglobin levels that are greater than normal limits, so this can affect the morphology of red blood cells. In this study, red blood cells were classified into 5 classes based on their shape, namely ellyptocytes, ovalocytes, schistocytes, tear dops, and normal using several digital image processing techniques and a backpropagation neural network system with 20 parameters which were the result of a combination of features, namely morphology, texture , and geometric invariant moments. The highest training accuracy rate is 93.94% and the highest test accuracy rate is 88% at a learning rate of 0.6 with a total data amount of 390 data.
Polycythemia Vera (PV) is one of the categories of Myeloproliferative Neoplasms Syndrome which is characterized by an increase in the number of red blood cells (erythrocytosis) as well as an increase in hemoglobin levels that are greater than normal limits, so this can affect the morphology of red blood cells. In this study, red blood cells were classified into 5 classes based on their shape, namely ellyptocytes, ovalocytes, schistocytes, tear dops, and normal using several digital image processing techniques and a backpropagation neural network system with 20 parameters which were the result of a combination of features, namely morphology, texture , and geometric invariant moments. The highest training accuracy rate is 93.94% and the highest test accuracy rate is 88% at a learning rate of 0.6 with a total data amount of 390 data.
Download
book
BibTex
Latex, Jabref
cloud_download
Original Resource
url resource Institution
