An Unobtrusive Computerized Assessment Framework for Unilateral Peripheral Facial Paralysis

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Computer Vision

Background:

• Unilateral peripheral facial paralysis (UPFP) assessment is crucial for neural rehabilitation.
• Current House-Brackmann (HB) grading is subjective and varies between observers.
• An objective UPFP assessment method is needed.

Purpose of the Study:

• To develop and validate an objective method for assessing UPFP using webcam image sequences.
• To compare the proposed objective method with the subjective House-Brackmann (HB) grading system.

Main Methods:

• Webcam image sequences of 5 healthy and 27 UPFP subjects performing facial actions were captured.
• Facial regions and landmark points were identified using Haar cascade and supervised descent methods.
• Support vector machine (SVM) classifier trained on extracted facial features achieved objective UPFP grading.

Main Results:

• The proposed method achieved an exact match rate of 49.9% and a loose match rate of 87.97% with HB scores.
• After sample augmentation, the accuracy increased to 90.01%.
• The system demonstrated superior performance compared to previous reports.

Conclusions:

• An unobtrusive webcam setup can be used for objective UPFP assessment.
• The proposed framework shows encouraging results for automated facial nerve function evaluation.
• This objective method has the potential to enhance UPFP diagnosis and rehabilitation.