Ray-based extraction of 5D optical information using a uni-form light transformation framework.

We present a deterministic intuitively simple forward framework for extracting five-dimensional (5D) ray information-comprising spatial (x, y, z) and angular (θ, φ) patterns-from conventional focal stacks of 2D images, so-called uni-form light transformation (ULT). The proposed ULT transforms focal images such that only the optical contribution originated from a selected location remains consistent in shape and brightness across transformed images. This enables robust extraction of directional and depth-resolved structure, even under dense or scattering conditions. We validate the approach using a widefield fluorescence microscopy dataset of C. elegans, demonstrating improved structural preservation, angular occlusion interpretation, and viewpoint-resolved imaging compared to conventional deconvolution methods.