Research On The Evaluation Algorithm Of CPR With Image

CPR (cardio pulmonary resuscitation) has become the world’s most respectedand most widely popular first-aid technique and the skill training of it means a lot tothe emergency nursing teaching. As the development of medical technology andcomputer science, a large amount of implements using in CPR teaching and training,among which the most representative is the dummy man for medical training.However, the teaching system of cardiopulmonary resuscitate dummy man collectsthe breath data and the depth of the operator’s pressing and makes judgment to themovement of operator’s action,just to the point of imitating patients’ interiorphysical sign. But it ignores the correctness of the pressing posture, hand posture andvertical strength, thus can be clearly seen that the system is not enough exhaustiveand further complete should be taken. At the same time,as to the evaluation ofcardiopulmonary resuscitate,the effect of CPR operating is related to several aspectsincluding pressing depth,pressing position, rate of pressing,rate of breathing,vertical strength and correct hand posture. To solve the current problem ofcardiopulmonary resuscitate teaching system is to analysis and integrate the aboveelements through certain method, so that to conform a synthesized cardiopulmonaryresuscitate evaluation. Therefore, this article is about certain aspects of which thespecific tasks are as follows:The article is about the general design of cardiopulmonary resuscitateevaluation model including three parts: the quantified processing of pressing andbreathing data collection pressing of dummy man’s interior sensors, identifying thepressing posture,vertical strength and hand posture by the image sensors, andgetting the synthesize evaluation with each graphic sensor’s data.Among which,thefocus part and difficult point of this article are the identifying of hand posture andthe establishment of the synthesize evaluation model.As to the section of CPR hand posture recognition,this article references on astatic gesture recognition method and designs a gesture recognition method based on a combination of features.First of all,to obtain the operator’s hand pressing binaryimage,it makes use of the ellipse skin colored model and the corresponding graphicpreprocessing;In the feature extraction,a structural feature extraction algorithmbased on the convex hull and depression profile was proposed in this article,usinghand gesture feature such as the number of fingers, included angle of finger tips asthe contingent feature,and using the improved Fourier descriptors as the globalfeatures,thus forming the combination features of cardiopulmonary resuscitate handgesture;In the hand gesture recognition,according to the characteristics of each ofthe contingent features and global features,a quick recognition method which cangradually excluded was designed.At the last, this article made use of the templatematching algorithm based on Euler’s distance to recognition.Experimental resultsshow that the proposed method can effectively distinguish between right and wrongof the operator’s pressing hand posture.To the aspect of cardiopulmonary resuscitate comprehensive evaluation, tosynthesize each evaluation index data collected by the complex sensors,this articleproposed the idea of data fusion and the decision-level evaluation model based onsupport vector regression to make an objective and comprehensive cardiopulmonaryresuscitate evaluation of students. In order to improve the accuracy andgeneralization ability of the model, this article made use of the mixed kernelstructure to establish the support vector regression model and proposed a chaoticdifferential evolution algorithm to get the optimization parameters of mixed kernelSVR model, and then made the comprehensive evaluation of students’cardiopulmonary resuscitate using the improved support vector regression algorithm.Finally,verified the effectiveness of the proposed method by experiments.Theexperiment shows that the evaluation model established in this article can effectivelyfuse the integrate each index factor collected by the sensors and obtain thecomprehensive evaluation conclusion.This article has established a scientificevaluation system for cardiopulmonary resuscitation teaching system.