iMalaySign: Malaysian sign language recognition mobile application using Convolutional Neural Network (CNN) / Nurul Natasha Mohd Jalani and Zainal Fikri Zamzuri

Sign language is used as one of the medium interactions among the deaf community in all around the world. Therefore, in Malaysia we used Malaysian sign language for deaf community to use in order to interact with people. Alphabets of Malaysian sign language also used as to spell some words, names, or any sign. However, sign language is only known by deaf people or people who learn sign language. Due to this, communication will not be delivered to anyone who do not understand or know sign language. This is due to the reason of people think sign language is hard to learn and sign language takes time to be learn. To add, some people for some reason had refused to learn sign language. Thus, a mobile application of Malaysian sign language recognition can be delivered to help deaf people communicate with normal people in our society. The scope of this research is focusing on five alphabets which covered alphabet (A-E). The objectives of this research as to design a mobile application that can help people knowing the meaning of alphabets, to develop this project by using image recognition technique and evaluate the functionality and accuracy of the sign language pose through the application. The methodology used in this application development is Modified Waterfall Model and CNN technique as the image recognition technique to recognize the alphabets pose. This application is tested with functionality and accuracy testing. As a result, this application is well function without any errors and can recognize the alphabet poses with a good accuracy. For future research is recommended to improve this application by increase the scope of the Malaysian sign language together with recognizing movement in Malaysian sign language.