Designing and Implementation of Novel Ensemble model based on ANFIS and Gradient Boosting methods for Hand Gestures Classification
Published in Proceedings of the 11th International Symposium on Information and Communication Technology, 2022
Recommended citation: https://dl.acm.org/doi/abs/10.1145/3568562.3568598
Abstract
Communication through hand gestures has always been the primary method worldwide. There are numerous methods implemented for the classification of hand gestures. However, most successful attempts use Convectional Neural Network (CNN) based methods, which are high in resource consumption and computational complexity, leading to problematic implementation in low-resource platforms. This study proposes a lightweight, efficient, and effective hand gesture recognition method using a novel ensemble model. The proposed ensemble method used XGBoost, CatBoost, and LightGBM gradient boosting algorithms as the initial classification. Then the Adaptive Network-based Fuzzy Inference System (ANFIS) algorithm was used to boost the accuracy. The dataset used in this study contains seven hand gesture classes and 350 samples, while each sample was generated using 30 frames summation of a sequence. The results were evaluated using two methods such as 10-fold cross-validation and confusion matrix-based parameters. The proposed ensemble algorithm outperformed eXtreme Gradient Boosting (XGBoost), Categorical Boosting (CatBoost), and Light Gradient Boosted Machine (LightGBM) algorithms in average accuracy, precision, recall, and f1-score, having 99%, 0.99, 0.97, and 0.98.
Recommended citation: Namal Rathnayake, Tuan Linh Dang, and Yukinobu Hoshino. 2022. Designing and Implementation of Novel Ensemble model based on ANFIS and Gradient Boosting methods for Hand Gestures Classification. In Proceedings of the 11th International Symposium on Information and Communication Technology (SoICT ‘22). Association for Computing Machinery, New York, NY, USA, 283–289. https://doi.org/10.1145/3568562.3568598.