摘要
拇外翻产生各种疼痛以及不良症状,主要是由前足足底压力过高导致。临床医生往往通过主观判断结合影像学诊断足底压力异常。由于缺乏足底压力数据支持,医生诊断与实际病情可能不相符。使用高传感器密度的阵列式压阻传感器设计了足底压力采集系统,实现了足底压力的精准采集、实时显示与存储。引入局部压力峰值分布,辅助实现拇外翻患者病情精准评估。还通过Resnet18对足底压力数据进行识别与分类,与现有6种算法相比,分类准确率提升1.1%。所提出的足部辅助诊断系统在影像学基础上通过阵列中的局部压力分布辅助医生进行主观诊断,对于手术与治疗方案的选择以及提升拇外翻判别准确度至关重要。
Hallux valgus leads to various kinds of pain and adverse symptoms,mainly caused by high pressure on the plantar of the forefoot.Clinicians often diagnose abnormal plantar pressure through subjective judgment combined with imaging.Due to the lack of plantar pressure data,the doctor's diagnosis may not match the actual condition.Therefore,improving the accuracy of hallux valgus discrimination through plantar pressure distribution on the basis of imaging is an urgent problem for the diagnosis of this type of disease.In this paper,a plantar pressure acquisition system is designed using an array piezoresistive sensor with high sensor density,which realizes accurate acquisition,real-time display and storage of plantar pressure data.Local pressure peak distribution is introduced to assist in accurate assessment of hallux valgus patients.Moreover,Resnet18 is used to identify and classify the plantar pressure data.Compared with the existing six algorithms,the classification accuracy is improved by 1.1%.The proposed foot auxiliary diagnosis system assists the subjective diagnosis of doctors through the local pressure distribution in the array on the basis of imaging,which is very important for the selection of surgery and treatment plan and the improvement of the accuracy of discriminating Hallux valgus.
作者
刘畅
陈煜
冯立辉
及松洁
石锐
卢继华
LIU Chang;CHEN Yu;FENG Lihui;JI Songjie;SHI Rui;LU Jihua(School of Medical Technology,Beijing Institute of Technology,Beijing 100081;School of Beijing Institute of Technology Materials,Beijing 100081;School of photoelectricity,Beijing Institute of Technology,Beijing 100081;Jishuitan Hospital,Beijing,Beijing 100035;School of Beijing Institute of Technology IC&Electronics,Beijing 100081)
出处
《自动化与仪器仪表》
2023年第4期329-334,共6页
Automation & Instrumentation
基金
北京市朝阳区协同创新项目(CYXC2109)资助。