Fast and accurate spectrum sensing is one of the key functionalities in a cognitive radio (CR), targeted to exploit the unoccupied part of the licensed electromagnetic spectrum. The discrete Fourier transform filter bank (DFTFBs) employed in energy detection based spectrum sensing provides only fixed sensing resolution as it splits the wideband input signal into uniform bandwidth channels (frequency bands). In this paper, a modified DFTFB is presented which can be reconfigured with minimal overhead to extract channels of different bandwidths and thus enables variable resolution spectrum sensing. The proposed DFTFB consists of a prototype filter realized using the coefficient decimation technique for obtaining different passband widths. From the analysis of the architecture, it is clear that the reconfiguration over-head to obtain different sensing resolution in the proposed filter bank is only a few adders whereas reconfiguration of a conventional DFTFB involves reconfiguration of the prototype filter and DFT which is an expensive task. © 2010 IEEE.