Brain and pancreatic tumor segmentation using SRM and BPNN classification

As of late, to enhance the features of serviceability in medical clinic management, medical image processing plays progressive development in conditions of modus operandi and applications. Various techniques are used to diagnosis tumor parts in modern medical image processing with the rising demand in the respective field. In this paper, the detection of the brain tumor and pancreatic tumor using DBCWMF (Decision Based Couple Window Median Filter)algorithm, Statistical region merging (SRM), Cat Swarm Optimization and Scale-invariant feature transform (CSO-SIFT) extraction and classification through Back Propagation Neural Network (BPNN) is presented. DBCWMF works effectively in the preprocessing compared to Median and PGPD filter, segmentation done with SRM algorithm. After that, the feature selection techniques CSO and SIFT are used for detecting the part in tumor images which is affected and final classification through BPNN classification works effectively compared to ANN and AdaBoost classifier. The experimental tested on images from Medical Harvard School database and The Cancer Imaging Archive (TCIA) repository’s database.