Dynamic Evaluation Of Facial Paralysis Rehabilitation Based On Data Fusion

Patients with facial muscle and nerve damage will have a distorted expression,which seriously affects the patient’s daily life.For patients,it takes a lot of time and money to go to the hospital regularly for a long-term rehabilitation check.In addition,patients may not be able to show the best state due to psychological factors such as nervousness and fear when they face the doctor,which makes the assessment of rehabilitation status not accurate enough.For doctors,assessing the rehabilitation status of patients is a repetitive basic task that takes up a lot of outpatient time and reduces the number of patients admitted.In addition,at present,diagnosis and treatment are mainly based on the subjective judgment of doctors,and there is no objective evaluation system that can accurately evaluate the effect of patient training.Therefore,based on the needs of the rehabilitation surgeon of the Ninth People’s Hospital,we researched and designed a dynamic evaluation system for facial paralysis rehabilitation based on doctors’ data,which can provide doctors with facial recognition,facial symmetry quantitative calculation to assist doctors to evaluate their rehabilitation status and provide a facial symmetry evaluation model with dynamic expressions to make the evaluation results more objective and doctors’ treatment more efficient.The results achieved are as follows:First,a dense face keypoint detection model was constructed with machine learning method to accurately describe the feature points of key parts of the face for patients with facial paralysis.At the same time,a Faster-RCNN deep learning algorithm is used to construct an iris recognition model,which provides a reference for subsequent calculations.Then,based on the constructed model,a facial symmetry quantification system was completed.This system is based on the professional symmetry calculation scheme provided by the doctor,and calculates the symmetry for the relevant key points based on the face key point detection model.With an accurate corneal detection model,it can be used to construct a face coordinate system and a physical distance reference.Not only can the quantitative calculation of symmetry be more accurate,but it also provides a solution for cross-photo comparisons possible.Aiming at the above two parts,this research also developed a practical software with a graphical user interface.This software allows doctors to easily use the models and systems designed above to detect and obtain information on the patient’s face at any time.The system uses a server / client separation mechanism,so that users can use it without excessive hardware requirements.Finally,based on the above research,this paper designs a dynamic assessment system for facial paralysis rehabilitation based on doctor data and Sunny Brook assessment scale.Based on the data obtained by the above symmetry quantification system,features are selected as input according to the professional opinion of the doctor,and then the authoritative scoring data of the doctor is used as the output.A classification algorithm is applied to constructing an evaluation model.