Octave and Level Sampling-Based Medical Image Registration Using Maximum Dominance

Image registration is an important challenge in medical image analysis. Recently, medical image registration is considered as a valuable assistant for the medical experts. In standard image registration, the point of interest deals with the detection and selection of feature points. In this article, in addition to detecting and selecting the feature points, we have introduced motion vectors for detecting the specific object deviation by setting the threshold level. In the proposed method, convolutional filters are used which are more effective in searching the feature points. For filtering these features, Hessian differential operators, octave, and level sampling are used. For selecting feature points, interpolation parameters and non-maximum dominance of differential operators are used. Furthermore, block matching motion estimation algorithm is used in identifying the translational movement which is an essential measure in medical imaging, particularly to detect abnormalities related to loosening, subsidence, and anteversion.