Application of Gabor Representation to Military Problems
Final technical rept. 5 Jan 1990-4 Jul 1993
ATLANTIC AEROSPACE ELECTRONICS CORP GREENBELT MD
Pagination or Media Count:
In summary, the effort so far has proved in principle most of the supporting concepts, but has been insufficient to transition the work into the applications arena as yet. AAEC sees particular promise for this technology in certain applications areas, and is planning to propose effort in those areas. A key area is automatic target recognition ATR. Machine-aided recognition problems have the feature that searching for objects can be enhanced in circumstances where shape characteristics of the objects are partially known in advance, either through a prior knowledge or data-aided algorithms. For example, in signal analysis, the Gabor transform is particularly adept at finding features having a common envelope. To maximally exploit such a circumstance, extraction of the analysis window from he data looms important. Given a large body of data such as that often encountered in an ATR problem, use of the data to drive the analysis functions seems wise as a measure to cut the amount of blind search. especially in view of findings that allegedly more robust tools such as the Wigner distribution can create artifacts through nonlinear processing if not used carefully. The role for optimum Gabor windowing in this scheme is clear, and as a result it appears that the best way in which to continue the line of work discussed above is to do it within the context of an application area such as ATR. The research at the point where it could profit from the interaction with real data as an aid in algorithm development refinement. AAEC anticipates proposing a body of work of his nature as a logical follow-on to the work performed in both this contract and the cited SBIRs.
- Numerical Mathematics
- Computer Programming and Software