The circular harmonic expansion technique has been applied to the quantized amplitude-compensated matched filter for rotation-invariant recognition. The filter is an amplitude filter with three discrete grey levels. The performance of the proposed filter is compared with the circular harmonic binary filter which has amplitude values of zero and unity. Three different circular harmonic components, namely the zero-order, first-order and second-order, were used in the investigation. Computer simulations show that there is a trade-off between the Horner efficiency and the signal-to-noise ratio with the amplitude compensation technique. With the second-order circular harmonic, an improvement of 45 % in the signal-to-noise ratio was achieved with only a 12% loss in the Horner efficiency.
|Number of pages||5|
|Publication status||Published - 1998|