Answers are grounded in this paper and cite their sources.
This study presents a technical evaluation of the edge-based Spatial Frequency Response (e-SFR) method for determining the resolution of fibre-bundle endoscopes (fibrescope). The analysis focuses on the precision and reliability of measurements affected by the inherent honeycomb-like sampling pattern of fibre-bundle imaging systems. Direct application of the e-SFR method to honeycomb-patterned images resulted in significant edge-detection failures and large deviations. A modified implementation of the SFRMAT5 algorithm, incorporating filtration for improved edge detection, was evaluated using synthetic test images. Although the averaged results yielded errors below 10% within the 0.04–0.25 cycles/pixel range, the procedure was time-consuming and required repeated manual adjustments of the Region of Interest (ROI), along with defining data rejection criteria that remain subject to operator bias. Outside this range, inaccuracies reached up to 38%. The introduction of morphological pre-filtration improved robustness, enabling consistent SFR50 estimation with errors below 10% in the 0.02-0.06 cycles/pixel range in the test with synthesised images. However, application to real fibre-bundle endoscope data revealed limited reliability. Measured SFR50 values with the rejection-averaging method (range 0.0168-0.0210 cycles/pixel) fell within high-error regions identified in synthesised image assessment, resulting in weak correlation with the limiting resolution. While filtered approaches improved trend consistency (up to a correlation of r = 0.889), they did not ensure accurate resolution estimation. These results demonstrate that the standard e-SFR method is unsuitable for fibre-bundle endoscopes without substantial modification. Further work is required to establish a validated reference methodology and improve the absolute accuracy of resolution measurements for fibre-bundle imaging systems.
No takes yet. Share an insight, caveat, or question.
Wojciechowski et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: