The parallel complexity of template matching has been well studied. In this paper we present more work-eÆcient algorithms than the existing ones. Our algorithms are based on FFT primitives. We consider the following models of computing: PRAM and the hypercube.