Quantitative Analysis And Application Of Scoliosis Muscle Distortion Based On Ultrasound Images

The idiopathic scoliosis is as high as 4%,and the diagnosis and treatment of it need medical imaging technology.In the clinic,the most commonly used examination is a doctor measures the scoliosis angle(Cobb angle)in an X-ray orthophoto.However,there are three shortcomings:First,asymmetry of the Paraspinal Muscle is the main cause of scoliosis and exacerbation.Measurement and assessment of Paraspinal Muscle lesions is also one of the main basis for the diagnosis and rehabilitation of scoliosis,but the specific condition of the muscle cannot be understood through X-rays;The second is that X-rays have radiation and cause certain harm to the human body,which is not suitable as a frequent inspection method;The third is that there is an error in the manual measurement of the Cobb angle,and the results are not accurate.So,it is particularly necessary to study the lesions degree of scoliosis and the corresponding computer-aided diagnosis and treatment system.In this respect,this thesis studies the ultrasound images of paraspinal muscle in patients with idiopathic scoliosis.Specifically,Firstly,an ultrasonic muscle structure segmentation algorithm based on U-Net is proposed to improve the accuracy of muscle structure feature extraction;Then explore the differences of paraspinal muscle in patients with scoliosis from 16 angles of structure and texture;Experiments have found that muscle thickness,pennation angle,fascicle length,mean,integrated optical density and energy have significant differences and can reflect the degree of paraspinal muscle distortion.Based on these six parameters,a quantitative assessment model of the degree of distortion of the muscle state can be used to more intuitively show the difference of paraspinal muscle on both sides of patients with scoliosis more intuitively.It is expected to provide quantitative information and basis for the diagnosis of scoliosis and the formulation rehabilitation treatment programs in the clinic.Finally,an online analysis software system based on Python-Flask is designed and built.The software can extract 16 parameters of the paraspinal muscle’s ultrasound images online,according to the quantitative evaluation model of the degree of muscle distortion,calculate the longitudinal change trend of the paraspinal muscle between the upper and lower vertebral body of scoliosis;The muscle degree of difference between the spine and its changing trend;It can help clinicians to realize automatic qualitative and quantitative analysis of paraspinal muscle in patients with scoliosis,assist diagnosis,evaluate the therapeutic effect of the rehabilitation planIn summary,the study of paraspinal muscle s in patients with scoliosis ultrasound images is mainly divided into 4 aspects:(1)An ultrasound muscle structure segmentation algorithm based on U-Net is proposed;(2)Propose the main parameters that can quantitatively evaluate the muscle status;(3)Quantitative analysis of the degree of paraspinal muscle distortion in patients with scoliosis;(4)A software system for online analysis of the degree of distortion of the paraspinal muscles based on Python-Flask is designed.The work of this paper is expected to provide assistance in the diagnosis of scoliosis and the development of sports rehabilitation therapy.