We propose a circular scanning system combining a single pair of gamma-ray source and detector to minimize the size of the measurement system. Unlike computerized tomography using an X-ray tube, a low photon-flux occurs in that using gamma rays. To correctly process sparsely sampled and highly noisy data retrieved from the system, we used total-variation minimization for image reconstruction and compared its performance to that of other methods on images in 64 × 64 matrices with 180, 90, and 45 projections. The insufficient projection data retrieve higher quality images for iterative reconstruction, expectation maximization, and projection onto convex sets with total variation than for the conventional filtered back-projection.