Indonesian Sign Language Image Detection Using Convolutional Neural Network (CNN) Method
DOI:
https://doi.org/10.35585/inspir.v13i1.37Keywords:
SIBI, BISINDO, CNN, Neural Network, AccuracyAbstract
In Indonesia, there are two sign languages utilized by the deaf community, SIBI and BISINDO. Unfortunately, the majority of non-deaf individuals and deaf companions are not proficient in sign language. To address this communication gap, information systems can play a pivotal role in recognizing sign language speech. Recently, researchers conducted a study using the Convolutional Neural Network (CNN) algorithm to predict sign language for both SIBI and BISINDO datasets. The aim was to develop a model that could accurately translate sign language into written or spoken language, thus bridging the gap between deaf and non-deaf individuals. The research found that the CNN algorithm performed optimally on epoch 50 for SIBI with a testing accuracy of 93.29 %, while for BISINDO, it achieved the best result on epoch 40 with a testing accuracy of 82.32 %. These results suggest that the CNN algorithm has the potential to accurately recognize and translate sign language, thus improving communication between deaf and non-deaf individuals in Indonesia.
Downloads
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2023 Andreas Nugroho Sihananto, Ms. Erista Maya Safitri, Mr. Yoga Maulana, Mr. Fikri Fakhruddin, Mr. Mochammad Ervinda Yudistira
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.