Accuracy and Precision of Edge-Based Modulation Transfer Function Measurement for Sampled Imaging Systems

The real-time, edge-based, multidirectional modulation transfer function (MTF) measurement system has elucidated the MTF characteristics of sampled imaging systems. However, the accuracy and precision of edge-based MTF measurements are not well understood quantitatively because of the complicated shift-variant nature of the sampling process by the image sensor and the binning process in the edge-based method for producing a supersampled edge spread function. The edge-based MTF measurement is valid only when the edge positions imaged are assumed to be distributed uniformly relative to the sampling sites. This study demonstrates that precision, which is dependent on the edge angle, is counterintuitively high at a higher oversampling ratio of the binning regardless of the smaller pixel counts per bin even with camera noise, and that significantly high precision can be achieved by taking an ensemble average of MTFs computed over a single cycle of the binning phase, even at low oversampling ratios, with high accuracy through proper corrections to compensate for the attenuation of the MTF values due to the discrete processing of the edge-based method.