Research On Flat Foot Gait Quality Evaluation Based On Neural Network And Its Application

The foot is one of the most important sensory organs of the body.It provides the dynamic and static information of the brain body as well as the basic structure for maintaining the daily posture of the human body sitting,standing,and walking.Due to the aging of the population and the persistence of children's overweight,the incidence of flat feet has increased year by year.The actual incidence of flat feet(including asymptomatic and symptomatic)in adults is much higher than the statistics.As a common foot symptom,flat foot is a potential cause of sports injury.Patients may also have loose midfoot,external rotation of the hind foot,trauma and neuromuscular imbalance.Patients with severe arch collapse cannot walk for a long time.Causes damage to joints and spine.The diagnostic methods of flat feet have gradually developed in recent decades,and more and more diagnostic methods have been derived.Common methods may be misdiagnosed and cannot fully assess the damage caused by the patient's arch collapse.Therefore,this paper has eliminated and optimized the experimentally measured indicators,and summarized a relatively objective and complete flat foot gait quality assessment system that can quantitatively reflect the severity of the flat feet of the subjects.In addition,based on the experimental data,a flat foot gait quality BP neural network evaluation model is constructed.The model can be used as a reference for clinical diagnosis,and an evaluation method can be provided when customizing the personalized footwear to better achieve precise treatment and rehabilitation.The main work of this paper is as follows:First,the Zebris running platform and the walking test design in the Vicon environment.Subjects will undergo a preliminary treadmill test and a further Vicon environment walking test to collect time-space parameters,plantar pressure related parameters,kinetic parameters,and kinematic parameters of the two groups under normal conditions.Repeat the test to ensure the integrity and validity of the data.Second,test data analysis.First,the reproducibility of the indicators is screened,the indicators with good reproducibility are retained,and the indicators with poor reproducibility are discarded.Correlation analysis is performed on the measured parameters,and the correlation index with strong correlation is eliminated because it is duplicated with other indicators.Second,the difference analysis between groups was conducted.According to the severity of flat feet,the subjects were divided into two groups:normal group and flat foot.The differences between the two groups were analyzed.The remaining indicators were analyzed by ANOVA,and the indicators with large differences were retained.value.Finally,principal component analysis was performed to select a total of 10 principal component factors with eigenvalues greater than one.List its load index,load factor,and variance contribution rate.The subject's balance score is calculated based on the load index coefficient design formula in the load factor.Third,build an availability assessment BP neural network model based on experimental data.Taking the data obtained in the two experiments as the training samples,after determining the network structure,training function,etc.,the optimal prediction model is obtained.The learning ability and generalization ability of the BP neural network model are tested by verification experiments.