<![CDATA[The standarized Indonesian Sign System (SIBI) is one of the media that helps communication among the deaf and mute in a wider community. It is known that 211.889 Indonesians are persons with disabilities consisting of 6.5% (13.802) are deaf and 2.6% (5.580) are speech impaired. Many ordinary citizens do not understand sign language which becomes a limitation for communicating with the deaf or mute. This research develops an application named IBIS with real-time sign language translator system. The real-time sign language translator system is developed using the convolutional neural network method with Tensorflow Lite Model Maker as the development medium. Researcher used the convolutional neural network method as the accuracy is relatively high. The model is integrated into Android based application developed with Flutter framework. IBIS application development starts from system and interface design using the waterfall method. Furthermore, the system is implemented in accordance to the defined requirements. The model is integrated into the Android based application using tflite_flutter and tflite_flutter_helper plugin. After that, testing is carried out for IBIS application and object detection model. The test for application testing includes validation testing and usability testing. The validation test is carried out using the blackbox method with the results show that all functionalities is in accordance with the defined requirements. Usability test with System Usability Scale (SUS) method reached a value of 86 and fall into the acceptable category. Testing for object detection model is done by comparing the original class with the detected class. The accuracy test reached 88% for 15 classes.]]>
Copyrights © 2022
