| Autor: |
Nugroho, Asep, Rozaqi, Latif, Sya'ban, Yukhi Mustaqim Kusuma, Yusuf, Sharfiden Hassen, Lukodono, Rio Prasetyo |
| Zdroj: |
AIP Conference Proceedings; 2024, Vol. 3069 Issue 1, p1-9, 9p |
| Abstrakt: |
Non-verbal communication usually uses for human computer interactions. One of non-verbal communication uses sign language such as hand gestures. To understand the command through hand gestures, the machine should recognize every movement so it needs classifier algorithms. Common classifiers algorithms such as k-nearest neighbor, decision tree, and support vector machine are used to recognize American sign language gestures. This study provides comparison performance among three common classifiers. Based on performance assessment, the support vector machine has the best accuracy and precision performance compared to two other algorithms. It can achieve 81.11% accuracy and 83.54% precision. In another hand, the k-nearest neighbor gives 74.44% accuracy and 74.40% precision while decision trees deliver 74.44% accuracy and 75.30%. Overall algorithms that are assessed giving accuracy below 90%. The features extraction that is based only on the time domain is not enough to spot distinction especially for signal that has a quite similar pattern, although it is a different class. In order to achieve more than 90% accuracy deep convolution neural network could be proposed as a future works. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
