Development of facial expression detection model for stroke patients

Abstract

A stroke patient should be cared and treated closely since the patient cannot speak, communicate, and move the body when needed. In addition, the number of stroke patients is increasing in Thailand and around the world. Unfortunately, the number of medical staffs does not vary by the number of stroke patient. Thus, the aim of this research is to develop a facial expression detection model for stroke patients during their treatments. This research proposes the facial expression detection model for stroke patients from facial features, such as Interpalpebral Fissure (IPF), Palpebral Fissure Length (PFL), Palpebral Fissure Region (PFR), Inner Brow Raisers (IBR), Brow Lower (BL), Inner and Outer Lid Raiser (LR), and Lip Part (LP). These features are applied to develop the facial expression detection model using the decision tree algorithm. Furthermore, there are two factors that can determine the facial expression detection, which are gender and age. From the experiment, the proposed the facial expression detection model can identify normal situations and abnormal situations for stroke patients with 95% accuracy; the value of precision is 91% and the value of recall is 100%. This model can be assisted for raising patient’s safety.