Research And Implementation Of Bone Density Analysis Algorithm Based On Humerus Image

In the last few decades,along with the population aging process continues to accelerate,our country has gradually entered the aging society.osteoporosis has always been one of those potential threats to the aging population.The decline of patients' life quality and the increase of their mortality will certainly brings heavy burden to the society,beloved ones and themselves.By using computer science technique,electronic technology and experts system,image processing technology has acquired the capability to handle those complicated works which may take a lot of human resources and experts' work by original process.Medical image processing has got great progress and development in these years,and the corresponding technique is widely used in the auxiliary diagnosis.Using computer processing medical image aided diagnosis mainly has the feature of high precision,high speed and high reliability.On the basis of an extensive related literature,this thesis has conducted the through research of image processing on CT images.The thesis designed and implemented bone density analysis algorithm based on humerus Image and implementation.Using the ICETEK-DM6437-B as hardware platform,it achieved the preconditioning process of CT images on DSP platform.This thesis includes the image binarization processing,image segmentation,target extraction,mathematical morphology,Gabor wavelet transform and median filter processing.On the basis of predecessors' research,this thesis researched and implemented the bone density analysis algorithm based on area information in image processing method.The main achievements in this thesis are listed as follows:(a)Humerus CT image processing.A design of image processing system which applicable to the humerus CT image is designed in this thesis.It Completed humerus CT image in Matlab image processing software algorithm.Setting cross contrast to research the effect of global threshold method between local threshold method for dealing with the humerus CT image.In the order of Ostu algorithm,Kittler,the minimum comparing error method and Niblack algorithm,the thesis are made to choose the most suitable humerus CT image processing algorithms.This thesis combined Niblack algorithm,the Gabor wavelet(b)transform and mathematical morphology processing mode.And successively complete the task of target extraction of humerus CT image.(b)The design of the BMD evaluation algorithm.Combined with a large number of medical literature,this thesis summarized the calculation of bone mineral density in image processing.The measurement standard is given by the world health organization(WHO)for estimating the humerus bone density analysis,a set of feasible osteoporosis methods is given in this thesis.By combining different locations on the same sample of bone mineral density measurement,along with the specific numerical support,this thesis proceed the auxiliary diagnosis.Verified the calculation results,this algorithm have proved itself useful.(c)On the ICETEK-DM6437-B hardware,this thesis achieves the function of the system.By using C language,it achieves transplantation the software application in CCS.The software carried on DSP system platform can complete the fast and exact analysis of CT images,achieved rapid auxiliary diagnosis.In this thesis,a software is designed on the Matlab platform,and a hardware system is carried on the DSP platform.Through the tests of 10 groups of CT images,the experimental results showed that the design in this thesis achieved the rapid analysis of humerus bone mineral density and the purpose of auxiliary diagnosis.