Physics-prior-guided compensation for spatially varying degradation in confocal endoscopy

Field-dependent blur and frequency-response loss limit miniature confocal endoscopic imaging. We introduce a physics-prior-guided dual-branch residual network that combines the degraded image with normalized field coordinates and calibration-derived MTF, PSF, SNR, and intensity-response maps. For the representative peripheral resolution-target region, the peak-to-valley difference increased from 0.23 to 0.36, and image-derived MTF50 increased from 16.7 to 40.4 lp/mm. Across 196 patches extracted from the four displayed source images, the proposed method achieved an NRMSE of 0.1394 and an SSIM of 0.8264. The results demonstrate spatially adaptive compensation by integrating calibration-derived physical priors into a computational reconstruction framework compatible with the existing imaging system.