Analysis of foreground detection in MRI images using region based segmentation
R. Kotteswari, K. Sathiya · 2016
Foreground Detection is one of the critical parts in the field of Computer Vision that aims to identify changes in the image sequences and to separate the foreground image from their background image. It is an arrangement of systems that typically examine the video sequences progressively and are recorded with a stationary camera. To detect brain tissue at early stage, a robotized framework utilizing images caught from multispectral camera is proposed. This multispectral camera has point of interest of catching images at higher determination. This can be implemented using region based segmentation (RBS) algorithm. There are two primary stages; first stage involves segmenting the input image by using preprocessing, which is performed by Anisotropic diffusion filtering and region segmentation. Second stage includes classification process by using Support vector machine (SVM) classifier and their main contribution is sobel edge detection and feature extraction. Experimental analysis indicates that the proposed method performs on diverse dataset with specific accomplishment on testing scenes that contain unpredictable or multiple associated foregrounds. Moreover, the change in accuracy is accomplished with low computational cost and executed utilizing MATLAB.