Assessing 3D Holographic Reconstruction Quality through the Mean Local Peak Signal-to-Noise Ratio Metric

Evaluating the quality of three-dimensional holographic reconstructions presents significant challenges, as traditional metrics often fail to accurately represent the match between reconstructed images and the target images, especially in terms of local details. To address this issue, we introduce a novel metric, the mean local peak signal-to-noise ratio (MLPSNR), designed to quantify the quality of three-dimensional holographic reconstructions more effectively. Unlike global assessment approaches, MLPSNR concentrates on the focused local areas of the reconstructed image, leveraging depth information to reflect the average quality of reconstruction. Through experimental validation, we demonstrate the efficacy and superiority of MLPSNR in evaluating the holographic reconstruction methods for three-dimensional neuronal structures.