For the last 18 years, our group has been developing a variety of near-field-to-far-field transformations (NFFFTs) for predicting the far-field (FF) RCS of targets from monostatic near-field (NF) measurements. The most practical and mature of these is based on the reflectivity approximation, commonly used in ISAR imaging to model the target scattering. This image-based NFFFT is also the most computationally efficient because - despite its theoretical underpinnings - it does not explicitly require image formation as part of its implementation. This paper presents a formulation and implementation of the image-based NFFFT that is applicable to two-dimensional (2D) spherical and one-dimensional (1D) circular near-field measurement geometries, along with numerical and experimental examples of its performance. We show that the algorithm's far-field RCS pattern-prediction performance is quite good for a variety of frequencies, near-field measurement distances, and target geometries. In addition, we show that the predicted RCS statistics remain quite accurate under conditions where the predicted far-field patterns have significantly degraded due to multiple interactions and other effect.