Correlation filters for object detection in nonoverlapping background noise using a noisy reference image

Classical correlation filters for object detection and location estimation are designed assuming that the appearance and the shape of the target are explicitly known. In this work we assume that the target is given at unknown coordinates in a reference image corrupted by additive noise. Optimal correlation filters, with respect to signal-to-noise ratio and peak-to-output energy, for object detection and location estimation are derived. Two mathematical models of observed images are used; the additive noise model for the reference image and the non-overlapping background model for the input scene. Computer simulation results obtained with the proposed filters are presented and compared with those of common correlation filters.