Fast implementation of polynomial-based interpolators for biomedical image rotation and scaling

This paper proposes the use of a polynomial interpolator structure (the Farrow structure) which is efficiently realizable in hardware, for high-quality geometric transformation of 2-D and 3-D images. Polynomial based interpolators such as B-splines are exactly implementable in the Farrow structure framework. Using this structure, the advantage in computational complexity is demonstrated to be very significant, typically saving around 80% of the computational cost of online direct interpolation.