A novel approach for multimodality medical image fusion over secure environment

Abstract

Summary An emerging trend to deal with the issue of multimodality medical image fusion over the secure communication environment is required. Hence, this paper presents a human visual fusion algorithm for medical images such as computed tomography and magnetic resonance imaging which is based on the nonsubsampled shearlet transform (NSST) over the secure environment. Initially, the images are decayed through NSST into low and detailed highlights. The neighborhood aggregate of correlation‐based movement measures is proposed to intertwine the low frequency and sum‐modified‐Laplacian‐based fusion is utilized on detail subbands of NSST. After getting a fused image, a method noise thresholding approach is utilized to improve the accuracy of the proposed method. As a result, more accurate fused images is received. To measure the proposed method accuracy, the results of the fused images are also tested over the secure environment where encrypted/decrypted fused images are obtained using the random generator method. These decrypted fused images are also analyzed with original fused images with the proposed method as well as existing methods. From result analysis, it is observed that the proposed technique is better in holding a bone, calcification, cerebrospinal liquid, edema, and tumor subtleties of the source images and is more accurate than other existing fusion methods.

A brain tumor detection system using gradient based watershed marked active contours and curvelet transform

“…The continuous curvelet transform (CCT) is represented by polar coordinates r and θ along with x spatial variable and w frequency domain variable. In CCT the image is represented with various windows each at different scales and orientations 16 . The frequency window U j ( r , θ ) in the Fourier domain can be defined as 24 : …”

“…Those values can be considered as undesirably high demonstrating inefficient segmentation and feature extraction. As is known, CT is one of the multiresolution image analysis techniques and there are many studies which have used this technique in brain tumor detection 5,16‐18 recently. In Table 1, summary of existing work related to brain tumor detection is given 2 …”