Research On Scoliosis Detection System Based On Digital Image Processing

In recent years,the incidence of adolescent scoliosis has been increasing year by year,and it has become a serious disease affecting the physical and mental health of teenagers.If it is not detected and corrected early,the spine may become deformed.Severe scoliosis can cause cardiopulmonary dysfunction,even paralysis and death.At present,the scoliosis examination methods are limited to doctors’ manual screening and X-ray diagnosis.These inspection methods are not only cumbersome and inefficient,but also cause unnecessary radiation damage.These ways are not good to large-scale investigation of adolescent scoliosis.Therefore,it is of great practical significance to study a fast,accurate and scientific detection technology of scoliosis.The paper analyzes the system functional requirements and technical indicators in detail,and presents a scheme for scoliosis detection system based on digital image processing.A scoliosis image detection algorithm was proposed according to the scheme,a hardware detection system was built,an embedded software program was written,a PC host software was designed,and wireless communication between the two was achieved.According to the characteristics and detection needs of the back image,a back image localization extraction algorithm was proposed,which successfully extracted part of the spine image from the back image.Two kinds of spine contour feature point extraction algorithms were proposed,one based on the outer contour of the back spine and the other one Based on grayscale feature differences.Two methods for evaluating scoliosis are proposed,one is based on the asymmetry degree of body feature asymmetry,and the other is based on the fitted spine contour curve to calculate the Cobb angle.A two-dimensional spine model was pre-built.According to the system function requirements,TI ’s TMS320DM642 chip was used as the core processor to design the image processing hardware platform for scoliosis detection system.The design scheme of the main modules of the system hardware circuit,the design flow of the embedded software function modules,and the PC host Interface design of each function module of computer software.The experimental verification scheme of scoliosis detection system is designed,and the relationship between scoliosis and asymmetry degree of body characteristics is studied.Fisher discriminant method is used to classify the scoliosis model based on the experimental results.The results show that the accuracy of negative recognition is 100%,and the accuracy of positive recognition is more than 85%,so as to realize the preliminary discrimination and classification of scoliosis.Furthermore,the measurement results of the standard side bending angle of the X-ray film are compared with the measurement results of the system,and the results show that the system is effective and feasible.In this paper,a fast,accurate,non-destructive and radiation-free detection and evaluation of scoliosis based on digital image processing is realized,which provides a low-cost and high-efficiency solution for scoliosis detection.It has important practical significance for the general survey of scoliosis and the timely detection and diagnosis of the disease.