¹Brawijaya University, 8 Veteran Road Malang, Indonesia.
²National Institute of Technology (ITN Malang), KM 2 Karanglo Malang, Indonesia.

---

Abstract—Hand gesture recognition is an interesting and challenging study, especially for human interaction with computer. There are many method can used for hand gesture recognizing such as Random K-NN, Tree K-NN, Fuzzy K-NN. K-Nearest Neighbor method is interesting to be examined. While the weighting method Simple Multi Attribute Rating Technique (SMART) can be used to improve classify accuracy result. That method known very simple to use start from determine criteria of weighting, normalization, utility value and recommendation. Thus, in our research, we combine SMART weighting and K-NN classification for alphabet sign language recognition to help disabled to communicate with each other. The simulation result, we have the average accuracy 94%, 95% and 96% for rather dark, normal and rather bright respectively.

Index Terms—K-NN, recognition, SMART.

[PDF]

Cite: Fitri Utaminingrum, I Komang Somawirata, Gilbert Dany Naviri, "Alphabet Sign Language Recognition Using K-Nearest Neighbor Optimization," Journal of Computers vol. 14, no. 1, pp. 63-70, 2019.