A Cloud-Based Training System For Facial Paralysis Rehabilitation And Evaluation Model

With the rapid development of information technology,many medical fields,including post-operative rehabilitation,require advanced technologies such as big data analysis and intelligence for a higher efficiency.In the field of facial paralysis,traditional training methods require a large amount of medical resources,and patients are difficult to insist on training.Moreover,there is a lack of an evaluation standard,and doctors need to do a lot of repetitive work.Due to the growing number of patients and limited medical resources,how to use intelligent,and automated methods to guide patients to do rehabilitation training and establish an evaluation model is an urgent problem to be solved.In this paper,a complete cloud-based facial rehabilitation training system and evaluation model were designed and developed.The system not only provides efficient rehabilitation training for patients,but also offers automatic traceability,professional facial training for patients and doctors.The research achievements are as follows:First of all,a client was designed and developed which can guide patients to carry out efficient rehabilitation training.Taking into account the training requirements of different facial parts,different training modes for different facial areas are designed.At the same time,in order to establish the follow-up evaluation model,the system will automatically collect a variety of information such as training videos,facial images,facial text data,etc.,which realizes automatic collection and storage of various types of information.In order to ensure the uniformity of training and the authenticity of the collected data,a measurement and calibration module is designed to ensure the real and effective data.Secondly,in order to trace and evaluate the patient's rehabilitation status in time,this paper also builds an intelligent cloud platform.It can provide patients and doctors with timely access to the interface of rehabilitation status.The information processing framework can automatically classify and store different types of data.Compared with the traditional face-to-face evaluation method,the establishment of the cloud platform greatly improves the efficiency of the patient's evaluation.Finally,based on the real data,an axis-based and point-based facial feature selection methods are proposed,which can effectively measure the patient's facial information,including asymmetry and offset index.From the static and dynamic perspectives,all aspects are considered and evaluated.And then,an evaluation model is established by using machine learning and cross-validation.The accuracy rate of the evaluation model is high enough to be used in clinical.