Unrolled Primal-Dual Networks for Lensless Cameras
This work addresses image quality issues in lensless cameras, which is incremental as it builds on existing primal-dual optimization frameworks.
The paper tackled the problem of inaccurate image reconstruction in lensless cameras by learning a supervised primal-dual method, resulting in a +5dB PSNR improvement over methods that do not correct for model errors.
Conventional image reconstruction models for lensless cameras often assume that each measurement results from convolving a given scene with a single experimentally measured point-spread function. These image reconstruction models fall short in simulating lensless cameras truthfully as these models are not sophisticated enough to account for optical aberrations or scenes with depth variations. Our work shows that learning a supervised primal-dual reconstruction method results in image quality matching state of the art in the literature without demanding a large network capacity. This improvement stems from our primary finding that embedding learnable forward and adjoint models in a learned primal-dual optimization framework can even improve the quality of reconstructed images (+5dB PSNR) compared to works that do not correct for the model error. In addition, we built a proof-of-concept lensless camera prototype that uses a pseudo-random phase mask to demonstrate our point. Finally, we share the extensive evaluation of our learned model based on an open dataset and a dataset from our proof-of-concept lensless camera prototype.