Browsing by Author "Mise, Pelin"
Now showing 1 - 1 of 1
- Results Per Page
- Sort Options
Conference Object Spine Posture Detection for Office Workers With Hybrid Machine Learning(IEEE, 2023) Öke, Deniz; Çakar, Tuna; Yıldız, Ahmet; Mise, Pelin; Terzibaşıoğlu, Aynur MetinThis study aims to detect bad spine posture using an al-ternative approach that doesn't rely on deep learning or excessive energy. The goal is to improve accuracy and effectiveness without disrupting workflow. A custom dataset was created, numerical inferences were made from posture values, and a hybrid approach using Light Gradient Boosting achieved a 96 % success rate.
