Open AccessOptical character recognition and long short-term memory neural network approach for book classification by librarians
Author(s) -
YD Rosita,
YN Sukmaningtyas
Publication year - 2020
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1567/3/032034
Subject(s) - computer science , dewey decimal classification , character (mathematics) , search engine indexing , cataloging , index (typography) , term (time) , cover (algebra) , information retrieval , subject (documents) , word (group theory) , font , library of congress classification , optical character recognition , artificial neural network , decimal , world wide web , artificial intelligence , library classification , image (mathematics) , linguistics , arithmetic , mechanical engineering , philosophy , physics , geometry , mathematics , quantum mechanics , engineering
The book is classified by librarians that use Decimal Dewey Classification (DDS) System. It is used for cataloging and indexing books. DDC has three divisions, a ten, a hundred, and a thousand. The book subject is reflected in each division. Commonly, to know the book content, librarians read the book title. Then, they identify the book index in DDC system. Nevertheless, it requires more time. To read the book title, Optical Character Recognition (OCR) aids them to get the book title efficiently that convert the image of the book cover into the text-editable. Librarians use a web camera to scan the book cover, especially the book title area. There are three steps for pre-processing, the lowercase changing, the useless word removing, and tokenizing. To detect the book categories, Long Short-Term Memory Neural Network is good implemented in this research. It is almost used for text classification. In this research, It gives high performance that achieves more than 92% accurately.