<![CDATA[Communication is an important aspect of social life, including for deaf people who rely on sign language as their primary means of interaction. However, the limited understanding of sign language in society, especially BISINDO (Bahasa Isyarat Indonesia), is often a barrier to effective communication. This study aims to develop a simple sign language translator system using TensorFlow and Keras-based machine learning technology, as well as image processing support from OpenCV. The system is trained to recognize six basic signs in BISINDO, namely “Halo”, “Nama”, “Saya, “A”, “D”, and “I”, using hand images as training data. Each word is represented by 20 images with variations in horizontal position. The test results show that the system is able to recognize signs with fairly good accuracy under certain conditions, but still has difficulty when faced with variations in hand positions that are different from the training data. This study shows the potential of sign language recognition technology in supporting inclusive communication, as well as the importance of data enrichment and variation in viewpoints in the model training process so that the system can function more optimally in various real conditions.]]>
Copyrights © 2025
