Monomodal Brain Image Registration Using Fast Walsh Hadamard Transform

Image registration is of great significance to medicine and remote sensing, so a lot of techniques have been developed within the context of one or the other discipline. This paper suggests a method for medical image registration using Fast Walsh Hadamard transform. This algorithm can be used to register images of the same or different modalities using CT or/and MRI images. Each image bit is lengthened in terms of Fast Walsh Hadamard basis functions. Each basis function is a notion of determining various aspects of local structure, e.g., horizontal edge, corner, etc. These coefficients are normalized and used as numerals in a chosen number system which allows one to form a unique number for each type of local structure. Performance evaluation is done based on the indexes Mutual Information, Correlation Coefficient and the Elapsed time for registration. The experimental outcomes show that Fast Walsh Hadamard transform achieved good results than the conventional Walsh transform in the time domain. In addition Fast Walsh Hadamard transform carried out medical image registration consuming less time.