total Variation (TV) methods are very effective for recovering "blocky", possibly discontinuous, images from noisy data. A fixed point algorithm for minimizing a TV-penalized least squares functional is presented and compared to existing minimization schemes.