Institusion
Institut Teknologi Perusahaan Listrik Negara
Author
Yusmira, Ocfien
Haris, Abdul
Prathama, Muhammad Fadli
Subject
Teknik Informatika
Datestamp
2022-09-20 02:28:35
Abstract :
Every human being has a different character from one another. There are many ways to recognize characters, one way to analyze handwriting. Different forms or patterns of handwriting are a reflection of the brain because they are ideas of thought that can originate from mood, emotional or physical health. The study of a person's character based on handwritten patterns or forms is called graphology. With the existence of graphology can be used as a means of an agency that will recruit employees or form a management team. Therefore, in this research, a system graphology application was built using Convolutional Neural Network (CNN) for character identification. CNN will classify the handwritten images into 10 categories based on the main parameters namely the left margin, the direction of the writing and the size of the writing. Based on the results of testing using the Confusion Matrix the results obtained from this study amounted to 50%.