A computationally efficient approach for template matching-based image registration

作者：Gaidhane Vilas H^*; Hote Yogesh V; Singh Vijander

来源：Sadhana : Academy Proceedings in Engineering Sciences , 2014, 39(2): 317-331.

DOI：10.1007/s12046-014-0230-0

• Normalized cross-correlation
• similarity measure
• eigenvalue
• covariance matrix
• template matching

Image registration using template matching is an important step in image processing. In this paper, a simple, robust and computationally efficient approach is presented. The proposed approach is based on the properties of a normalized covariance matrix. The main advantage of the proposed approach is that the image matching can be achieved without calculating eigenvalues and eigenvectors of a covariance matrix, hence reduces the computational complexity. The experimental results show that the proposed approach performs better in the presence of various noises and rigid geometric transformations.

• 出版日期2014-4