2026-04-19T08:54:45Z https://ipsj.ixsq.nii.ac.jp/oai

oai:ipsj.ixsq.nii.ac.jp:00217429 2025-01-19T15:30:05Z 1164:3027:10764:10847

Recognition of Finger Movement Disability Level of Post-Stroke Patient Based on Fugl-Meyer Assessment Using Surface EMG Recognition of Finger Movement Disability Level of Post-Stroke Patient Based on Fugl-Meyer Assessment Using Surface EMG Adhe, Rahmatullah Sugiharto Shuhei, Tsuchida I, Putu Alit Pawana Tsutomu, Terada Masahiko, Tsukamoto Adhe, Rahmatullah Sugiharto Shuhei, Tsuchida I, Putu Alit Pawana Tsutomu, Terada Masahiko, Tsukamoto 学習と支援 Regaining finger movement function is thought to be more challenging during rehabilitation because its muscle complexity. Thus, Fugl-Meyer Assessment (FMA) is employed by a doctor to evaluate manually the disability level of the finger movement. This will lead to subjectivity and risk of mistake during assessment. Thus, a system capable of predicting the disability level is required to aid the doctor in making more accurate judgement. This study aims to recognize the finger movement disability level based on Fugl-Meyer Assessment. The EMG recorded from 4 patients when they performed 7 movements based on FMA and extracting the time domain feature values. SVM and Random Forest were employed in classifying the disability level of each movement. SVM classifier could obtain better output in movement 4 which was 91.67% of accuracy and 0.78 of f1 score. Regaining finger movement function is thought to be more challenging during rehabilitation because its muscle complexity. Thus, Fugl-Meyer Assessment (FMA) is employed by a doctor to evaluate manually the disability level of the finger movement. This will lead to subjectivity and risk of mistake during assessment. Thus, a system capable of predicting the disability level is required to aid the doctor in making more accurate judgement. This study aims to recognize the finger movement disability level based on Fugl-Meyer Assessment. The EMG recorded from 4 patients when they performed 7 movements based on FMA and extracting the time domain feature values. SVM and Random Forest were employed in classifying the disability level of each movement. SVM classifier could obtain better output in movement 4 which was 91.67% of accuracy and 0.78 of f1 score. 情報処理学会 2022-03-07 eng technical report https://ipsj.ixsq.nii.ac.jp/records/217429 2188-8760 AA1221543X 研究報告ヒューマンコンピュータインタラクション（HCI） 2022-HCI-197 10 1 8 https://ipsj.ixsq.nii.ac.jp/record/217429/files/IPSJ-HCI22197010.pdf application/pdf 1.4 MB 2024-03-07