<![CDATA[Communication is the process of delivering information in social life. However, people with hearing impairments have difficulty exchanging information because of their limitations. The Indonesian Government's solution to this problem is the standardization of the Indonesian Sign Language System (SIBI). Previous research has been carried out related to SIBI alphabet sign recognition with flex sensor and MPU6050 using the Naive Bayes method. However, from 130 tests, the research only produced an accuracy about 43.85%. The problem is due to the limitations of the flex sensor orientation reading accuracy. From these, development research was carried out to improve the accuracy of SIBI sign recognition. From these, development research was carried out to improve the accuracy of SIBI sign recognition. In this study, there were 6 units of MPU6050 that functioned as finger and opisthenar orientation readers. Sign recognition using the Naive Bayes method based on the sensor orientation reading. The research produces data suitability between the controller and the Android application is 100% that the application can be used as a data media representation, Mean Absolute Percentage Error (MAPE) of the sensor readings at 1,471% from 24 tests, classification accuracy rate at 92.31% from 78 tests, and the average processing time at 20 ms from 20 tests.]]>
Copyrights © 2021
