The front propagating problems is an important area of application of Level Set Methods (LSMs) wh... more The front propagating problems is an important area of application of Level Set Methods (LSMs) where moving front plays a vital role in evolution of process. In traditional LSMs, the Level Set Function (LSF) defines the evolution of boundary. But it does not apply to the object whose boundary we've to calculate after some time interval i.e. the LSF becomes 'irregular'. As a remedy to this, we've to redefine the LSF so that it'll give the explicit result. The costly re-initialization is a method to do this. But this causes area loss which is one directional. Many ideas were proposed to avoid re-initialization process like Chopp's methods, GDRLSE1, GDRLSE2, and GDRLSE3. But they are not applicable for both the LSMs i.e. PDE based and variational one. We are proposing reaction diffusion (RD) based method which will avoid all the disadvantages of all methods mentioned above, will be stable. There is no need to go for re-initialization of LSF. We'll add a diffusion term into Level Set Evolution (LSE) equation. To have a stable numerical solution of RD based LSE, we'll use two step splitting method (TSSM) which will first iterate the LSE equation and then solve diffusion equation. Solving diffusion equation regularizes the LSF to ensure stability, thus avoiding costly re-initialization completely. The method is applicable for PDE based LSM as well as variational LSM. It gives very good results in terms of high boundary anti-leakage and noise immunity.
Based on interpolation of low frequency sub band images obtained by discrete wavelet transform (D... more Based on interpolation of low frequency sub band images obtained by discrete wavelet transform (DWT) and the input image, the brain tumor detection is obtained by using Haar wavelet transform.Both input image and database image is decomposed into different subbands by using DWT. Interpolation of low frequency subbands as well as input image is done. Database image is also decomposed by using Haar wavelet transform by two level and this database image is compared with the input image by using Mutual information principle. Same information present in both of images is used for image registration which is discarded and remaining information is considered that is nothing but Brain tumor.
The front propagating problems is an important area of application of Level Set Methods (LSMs) wh... more The front propagating problems is an important area of application of Level Set Methods (LSMs) where moving front plays a vital role in evolution of process. In traditional LSMs, the Level Set Function (LSF) defines the evolution of boundary. But it does not apply to the object whose boundary we've to calculate after some time interval i.e. the LSF becomes 'irregular'. As a remedy to this, we've to redefine the LSF so that it'll give the explicit result. The costly re-initialization is a method to do this. But this causes area loss which is one directional. Many ideas were proposed to avoid re-initialization process like Chopp's methods, GDRLSE1, GDRLSE2, and GDRLSE3. But they are not applicable for both the LSMs i.e. PDE based and variational one. We are proposing reaction diffusion (RD) based method which will avoid all the disadvantages of all methods mentioned above, will be stable. There is no need to go for re-initialization of LSF. We'll add a diffusion term into Level Set Evolution (LSE) equation. To have a stable numerical solution of RD based LSE, we'll use two step splitting method (TSSM) which will first iterate the LSE equation and then solve diffusion equation. Solving diffusion equation regularizes the LSF to ensure stability, thus avoiding costly re-initialization completely. The method is applicable for PDE based LSM as well as variational LSM. It gives very good results in terms of high boundary anti-leakage and noise immunity.
Based on interpolation of low frequency sub band images obtained by discrete wavelet transform (D... more Based on interpolation of low frequency sub band images obtained by discrete wavelet transform (DWT) and the input image, the brain tumor detection is obtained by using Haar wavelet transform.Both input image and database image is decomposed into different subbands by using DWT. Interpolation of low frequency subbands as well as input image is done. Database image is also decomposed by using Haar wavelet transform by two level and this database image is compared with the input image by using Mutual information principle. Same information present in both of images is used for image registration which is discarded and remaining information is considered that is nothing but Brain tumor.
Uploads
Papers by Sushil Sirsat