Abstract:
Synthetic Discriminant Function (SDF) has been the main area of focus for many
researchers. It has been applied successfully in different applications. This filter was designed
to overcome the limitations of Matched Filter and Phase Only Filter. The main idea behind
SDF design is to include the expected distortions in the filter design. This technique improved
the immunity to the distortions. SDF filter is computationally very intensive algorithm when
implemented in MATLAB. Its real-time implementation with the specialized hardware is
essentially required in order to achieve high speed. It is an optical technique, which is
digitized.
This research work is focused on the hardware design of Synthetic Discriminant
Function Bandpass Difference of Gaussian Filter. The objective is to provide faster and
cheaper solution to the object recognition problem. This work pays special attention to the
trade-off among area, cost, performance and precision aspects. Hardware design of SDF filter
is the first step towards designing fast and efficient hardware for correlation filters with
distortion invariance.
