Pyramidal Multiphase Level Set Framework for Brain Tumor Detection Using Generalized Technique of Painting Background
Gang Zheng王惠南 · Transaction of Nanjing University of Aeronautics and Astronautics · 2007
The pyramidal multiphase level set framework (PMLSF) based on the technique of painting background (TPBG) and the Chan-Vese model can detect multiple objects on a given image. However, the boundaries of the sub-object obtained by PMLSF-TPBG are not variable since a special color parameter is used in TPBG. To solve the problem, a new technique utilizing a varying parameter is proposed to ensure that PMLSF is effective for the detection of the desired boundaries of the sub-object. The interval of the variable color parameter is proved and the effects of the parameter are also discussed. Experimental results for the brain tumor detection show that different boundaries of the brain tumors can be detected with different color parameters. It is especially useful for clinical diagnoses.