Development and AI-Based Evaluation of a Joint-Level Movement Dataset for Orthopedic Hand Assessment
TOK 2026 Otomatik Kontrol Ulusal Toplantısı, Ankara, Türkiye, 8 - 10 Ekim 2026, cilt.1, ss.676-681, (Tam Metin Bildiri)
- Yayın Türü: Bildiri / Tam Metin Bildiri
- Cilt numarası: 1
- Basıldığı Şehir: Ankara
- Basıldığı Ülke: Türkiye
- Sayfa Sayıları: ss.676-681
- Açık Arşiv Koleksiyonu: AVESİS Açık Erişim Koleksiyonu
- Dokuz Eylül Üniversitesi Adresli: Evet
Özet
Tracking and interpreting finger and hand movements provides valuable information that can help clinicians assess the recovery progress of patients undergoing physical therapy. In this study, a two-stage data collection process was carried out for the assessment of basic orthopedic hand problems, using two generations of a resistive flex sensor data glove. In the first stage, a dataset of 3096 samples covering sixteen movement classes was collected with a commercial reference glove, and the prototype signals were compared with the same system. In the second stage, a fifteen-channel prototype measuring three joints of each finger independently was designed, and a dataset of 2268 samples covering ten movement classes was collected. Both datasets were obtained from healthy volunteers under ethics committee approval, and their separability was evaluated through classification experiments. The datasets were subsequently used for orthopedic hand movement recognition.