High resolution wideband near-field imaging requires wideband signals and large array apertures. Hence, a large amount of data needs to be processed in the image formation step. The standard delay and sum beamformer scales linearly with both output size and number of measurements resulting in high computational load. We propose a novel approach for image formation reducing the numerical complexity based on the non-equispaced fast Fourier transform (NFFT). The conventional delay and sum beamforming algorithm is reformulated such that the computational efficiency of the NFFT is exploited. Experimental results from a through-the-wall radar imaging system show a good improvement of the speed whereas the introduced error is kept very low.