From simulation to reality: robust image quality assessment and calibration error compensation in 3D display systems

Objective image quality assessment metrics (IQAs) play a critical role in both 2D and 3D display technologies. However, existing studies largely rely on open-source databases or simulated images, with limited empirical evaluations involving actual 3D image comparisons. Moreover, little research has explored the potential of IQAs to assist in correcting calibration errors in 3D display systems. In this study, we apply a range of widely used 2D IQAs to both simulated and real-world 3D image quality evaluations. Six representative metrics—selected based on their performance, computational efficiency, and industry relevance over the past two decades—are employed to assess image quality in 3D displays. By deliberately introducing deviations into the calibration parameters, we analyze the impact on perceived image quality and demonstrate that IQAs can effectively detect and compensate for minor calibration errors in 3D systems. Simulation and experimental results show that traditional metrics such as PSNR and SSIM underperform, while CW-SSIM and GMSD provide moderate results. In contrast, FISM and MDSI exhibit superior performance and robustness. These findings support the use of objective IQAs as a practical tool for assisting in the calibration and performance optimization of glasses-free 3D display devices.