Physical disorders are considered to be the most severe disability in patients with hemiplegia after stroke. Currently, most studies have used motion feature extraction methods and machine learning-based methods to evaluate the functional degree of post stroke in hemiplegic patients. This research collected feature data from patients under diverse experimental conditions and then fed them into different machine learning classifiers. However, few studies have compared which classifiers and experimental condition could achieve more precise assessments in a specific condition. In this paper, we compared the accuracy of four different classifiers in a conservative motion recognition method. A motion sensor was used for monitoring the upper limb action, and four conservative machine learning classifiers, which map the features to Fugl-Meyer scale, were chosen for comparison. Ten post-stroke hemiplegic-simulated subjects performed a group of predefined actions, and these motion data were used to generate a group of features reflecting the information of each predefined action. We input the features into four classifiers to generate corresponding classifiers. With the Support Vector Machine classifier, prediction accuracy at 97.79% was achieved in the experiment data, which outperformed previous reports. In conclusion, Support Vector Machines perform better than the other three classifiers in the assessment of the degree of post-stroke hemiplegics. It is encouraging that results have been generated with the proposed assessment method in this exploratory study.

A rapid and objective assessment of the severity of facial paralysis allows rehabilitation physicians to choose the optimal rehabilitation treatment regimen for their patients. In this study, patients with facial paralysis were enrolled as study objects, and the eye aspect ratio (EAR) index was proposed for the eye region. The correlation between EAR and the facial nerve grading system 2.0 (FNGS 2.0) score was analyzed to verify the ability of EAR to enhance FNGS 2.0 for the rapid and objective assessment of the severity of the facial paralysis. Firstly, in order to accurately calculate the EAR, we constructed a landmark detection model based on the face images of facial paralysis patients (FP-FLDM). Evaluation results showed that the error rate of facial feature point detection in patients with facial paralysis of FP-FLDM is 17.1%, which was significantly superior to the landmark detection model based on normal face images (NF-FLDM). Secondly, in this study, the Fréchet distance was used to calculate the difference in bilateral EAR of facial paralysis patients and to verify the correlation between this difference and the corresponding FNGS 2.0 score. The results showed that the higher the FNGS 2.0 score , the greater the difference in bilateral EAR. The correlation coefficient between the bilateral EAR difference and the corresponding FNGS 2.0 score was 0.9673, indicating a high correlation. Finally, through a 10-fold crossvalidation, we can know that the accuracy of scoring the eyes of patients with facial paralysis using EAR was 85.7%, which can be used to enhance the objective and rapid assessment of the severity of facial paralysis by FNGS 2.0.