Font Size: a A A

Casas Bone Segmentation And Positioning Research

Posted on:2006-04-05Degree:MasterType:Thesis
Country:ChinaCandidate:J G YinFull Text:PDF
GTID:2204360155465345Subject:Pattern Recognition and Intelligent Systems
Abstract/Summary:PDF Full Text Request
The skeletal age score mainly aims at classifying, scoring of the growth of the youth bones, and get the age of the growth of bones. This method includes clear conception of quantification, and enhances the precision of the calculation of bones age. Now, it has been become more and more popular in medicine area, such as predicting the height of children after grown, analyzing the development of the growing, diagnosing and ward shipping illnesses and treatments of internal secretion and development of growing, which effects tremendously on studying the growing of human beings. The traditional skeletal age score method is depend on the subjective observation of doctors, differing from different doctors, also difficult for people to estimate many various bones in the same time. The score is affected by doctors greatly. With CHN standard, we need to score more than one bone. It will take too much time if processing by human, and can' t be processed in batch. For the automation of the skeletal score, firstly we need to extract the feature bones from X-ray pictures, so as to offer the input for the next step. And the work researched in this paper is mainly focus on this area.The input of the CASAS is hand-wrist radiographs of enfant, before we could measure, classify and score, we must extract the feature bones which will be processed at first. In this paper, we used the local dynamic threshold method to decrease the effect of the uneven lighting, and make use of the character of the metacarpus highlight appearance to calculate the local threshold.This paper is focus on the areas as follow:1) Bring forward a new method of segment bones from hand-wrist radiographs, making use of the gray histogram of bones, muscleand background in hand-wrist radiographs taking on three different distribution;2) Take advantage of the transcendent knowledge of the structure hand-wrist in picture, and by the methods of morphology, intensity projection etc. We can mark the location of the feature bones, and discrete these bones from pictures by rotation and copy;3) The projection map of wrist has a clear increment between metacarpus and radius, so this position in contour of hand-wrist will has the greatest value of curvature. We can use this character to mark the position separating metacarpus and radius;4) In this paper we developed a test platform under Visual C++ platform, to test the algorithms provided in the paper, and observe result from experiments.The experiments indicates that the method can obtain good results in segmentation and locating of feature bones, and can establish a good foundation for next step of the feature parameters extraction.
Keywords/Search Tags:Skeletal Age, Image Segmentation, Intensity Projection, Curvature Analysis
PDF Full Text Request
Related items