Research And Implementation Of Algorithm For Dynamic Lumbar Radiography Measurement

Lumbar diseases,such as lower back pain(LBP),are common chronic diseases in middle-aged and older individuals,and degenerative lumbar instability is considered the most important factor causing LBP,sciatica and other symptoms.Due to the complexity of lumbar movement and the lack of specificity in clinical manifestations,there is still much controversy about the pathogenesis and clinical diagnosis of lumbar instability.For degenerative lumbar instability,a standardized diagnostic criterion has not yet been established.However,the application of imaging measurement technology provides the possibility for assisting doctors in clinical diagnosis.The traditional X-ray measurement technology has too much radiation dose and can not provide continuous images.At present,in clinical practice,the hyper-extension and hyper-flexion displacements and rotation angles in the sagittal plane on an X-ray are typically used as criteria for determining whether lumbar spine instability is present,but it can not reflect the dynamic activity of the lumbar spine.Therefore,the continuous dynamic image sequences of moving objects at low doses were obtained by quantitative fluoroscopy(QF)in this paper,and the dynamic motion law of lumbar vertebrae were studied.This paper is mainly divided into the following aspects:(1)The research progress of lumbar spine was summarized.This paper introduces some vertebral positioning and tracking algorithms in previous research,and proposes the KLT algorithm and iterative template matching algorithm used in this paper in view of its shortcomings.(2)In order to track vertebral corners KLT algorithm is introduced.Harris feature point detection method is used to solve the problem of rotation and scale transformation.Hessian matrix is introduced to eliminate the points with inconsistent optical flow or big jump,and pyramid image of original image is generated to meet the requirement of low speed of KLT algorithm.Two simulation verification models are constructed to verify the feasibility and effectiveness of the algorithm.(3)Using iterative template matching algorithm to track vertebral corners.Firstly,a large template is selected to roughly match the approximate position of the vertebral body,and the template is updated every 50 frames to improve the success rate of the vertebral body matching;then,a small template is selected at each corner to further optimize the matching accuracy.The effectiveness of the model is verified by the same simulation.The simulation results show that the displacement error of KLT algorithm is less than 0.29 mm and the angle error is less than 0.3 degrees.However,the iterative template matching algorithm often fails to match.KLT algorithm has high tracking accuracy,fast calculation speed and better tracking effect than iterative template matching algorithm.A group of LBP volunteers' lumbar spine image sequences were collected,and two algorithms were applied to actual clinical data.It was observed that KLT-based method was better than iterative template matching method.KLT-based method could be used to assist doctors in the diagnosis of lumbar spine diseases.