Feature Information Extraction For 2D Projection Image

There are any different image effects in the medical image processing because of the different equipment and the computing method. Therefore,the interested regions are also different in the medical processing algorithm. Compared with the soft tissue or the other more complex part of body, the bone tissue is more concerned about the edge information. In addition, the traditional algorithm is generally complex, slow, and not suitable for medical image. Therefore we need a new algorithm which has some advantage such as strong adaptability,good convergence,and high performance to suitable for medical image and to improve computing speed.In the first, the attribute histogram is used to improve partition effectiveness as we can confirm division class. The threshold of algorithm is adapted to different interval in the actual process of segmentation.And then,the difference between traditional algorithm and improved algorithm should be judged by basic graph-theoretic criterion.Secondly,a fast single algorithm based on pixel feature is proposed in the paper. The single algorithm is achieved by the 2D pixel theory and the edge characteristics of the actual connection.So the primitive dates can be optimize by the single algorithm and it is a effective primed way to image connection.In the last,a edge linking algorithm based on least square method and pixel information is proposed in the paper.There is a effective way to reduce the error rates and to improve the computing efficiency.It can solve some related issues such as image basic type and connection trend,and a asymptotic plan is adopted to accomplish the connection.Finally, the feasibility and security of algorithm is analyzed in the paper,and the results shows that the algorithm is feasible and effective.