A Precise Facial Paralysis Degree Evaluation with Severity Classification Using Image Processing and Neural Network

Abstract

Facial nerve paralysis is a common disease, which involves the paralysis of any structures innervated by the facial nerve. Facial paralysis can be on single or both sides of the face but more commonly occur on a single side. This disease can lead to significant psychical and functional hurt to patients. Nowadays in medical field, many methods and facial nerve grading systems were developed for estimating the facial paralysis. However, these existing methods have some drawbacks in the facial paralysis degree estimation process. To alleviate such drawbacks, in this paper, a technique for facial paralysis degree evaluation with severity classification is proposed. The proposed technique is mainly composed of three phases. In the first phase, the facial feature points are acquired via Feature Point Selection Algorithm (FPSA) from the right and left sides of the face with different expressions. Based on these face feature point values, the facial paralysis degree are estimated in second phase. Moreover in third phase, the patient’s severity is classified by using a well-known AI technique called Feed Forward Back Propagation Neural Network (FFBNN). The proposed technique is implemented and the results are compared with existing facial paralysis degree evaluation technique. The result exhibits the effectiveness of the proposed facial paralysis degree evaluation technique in classifying the facial paralysis patients to their severity classes based on the estimated facial paralysis degree. Also, the proposed technique with FFBNN provides high accuracy results than the conventional degree evaluation method.
Copyright © 2013 Praise Worthy Prize - All rights reserved.