Closed retact closed 3 years ago
https://github.com/takutosoeda/paper-survey/blob/main/survey/ARL1/2021-04-24%20Hand%20Gesture%20Recognition%20Based%20on%20Semg%20Signal%20and%20Improved%20SVM%20Voting%20Method.md できるだけ少ない特徴量でできるだけ多くのジェスチャーパターンを認識するSVMアルゴリズムの提案。従来のvoting Methodとは異なり全てのサブセットトレーニングに対応している他分離可能性を高めた方法の提案がされている。これにより特徴ベクトル次元数は3元で12のジェスチャーを84%の認識精度を記録した。
https://ieeexplore.ieee.org/document/9236928