This thesis describes automatic algorithm for the segmentation of subcutaneous fat and visceral fat on computed tomography(CT).
This algorithm was applied to 267 slices of 33 people.
With the result which applies the algorithm, 209slices(78.27%) sho...
This thesis describes automatic algorithm for the segmentation of subcutaneous fat and visceral fat on computed tomography(CT).
This algorithm was applied to 267 slices of 33 people.
With the result which applies the algorithm, 209slices(78.27%) showed a good result and, the result which violates a with each other region appeared from 17slices(6.37%), 41slices(15.36%) was not good.
In this not good results, the subcutaneous fat is breaking from near the ribs bone or the vertebral, or the subcutaneous fat and visceral fat are sticking nearly.
From like this input data of not good results, the visceral fat comes to be divided into several parts, or the violation happened with the visceral fat direction.
For like this not good results, it added the menu in program it will be able to correct by manual operation it complemented.
If the algorithm which is proposed from this thesis is used, it will provide being more convenient, and will provide the ground data of the obesity diagnosis which is accurate.