A new shifted phase-encoded fringe-adjusted joint transform correlation technique is proposed for pattern recognition. The proposed technique utilizes a random phase mask to the reference image along with a phase-shifted version of the phase-masked reference image before forming the input joint image. The phase mask is also applied to the joint power spectrum before applying the inverse Fourier transform to yield the correlation output. This yields delta-function-like correlation peak intensity for each target object and ensures better utilization of the input/output plane space bandwidth product. The proposed technique can effectively detect targets in simulated as well as real-life input scenes without adjusting the system parameters. Simulation results are presented to verify the performance of the proposed technique. © 2004 Published by Elsevier B.V.