Breaking the Under‐Display Camera's Dilemma Between Diffraction and Pixel Density Using Incoherent Pupil Synthesis

ABSTRACT A high pixel density, along with the resulting low aperture ratio, leads to strong diffraction and severe image degradation in under‐display cameras (UDCs). Previous studies have explored modifying pixel structure and image retrieval algorithms. However, these approaches are limited by the many zero values (i.e., gaps) in the optical transfer function (OTF), which are intrinsically caused by a periodic, low‐aperture‐ratio pupil function. This study proposes to segment the camera pupil into incoherent sub‐apertures with independently configured pixel structures. At high pixel density, though substantial gaps appear in the OTFs of the sub‐apertures, these gaps are complementarily eliminated after incoherent synthesis. As a result, by adopting a low aperture ratio of 24% and using Wiener deconvolution, retrieved images corresponding to the synthetic pupil show a clear improvement in image quality (> 0.07 in MS‐SSIM) compared with those retrieved from a classic rectangular pixel layout. UNet‐based retrieval shows a similar improvement, which also benefits from the more enriched physical information produced by the gapless OTF.