Anti-Aliasing Snapshot HDR Imaging Using Non-Regular Sensing

A team from Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU) has published a breakthrough paper on arXiv detailing a novel approach to snapshot High Dynamic Range (HDR) imaging. The research, led by Teresa Stürzenhofäcker, addresses a critical limitation in current HDR capture: the need for multiple exposures or complex hardware setups, which fail in dynamic scenes with motion. Their solution is a sensor that physically extends dynamic range by combining two differently sized prototype pixels with varying light integration areas, all arranged in a non-regular pattern. This irregular layout is the key innovation, designed to mitigate the aliasing artifacts and loss of spatial resolution typically associated with increasing pixel size for better light capture.

The technical core involves a subsequent reconstruction process in the Fourier domain, where natural images can be sparsely represented, allowing the recovery of high-detail imagery from the sensor's raw data. Simulation results confirm the proposed sensor layout effectively acquires images with both high dynamic range and freedom from aliasing. This work, also accepted for presentation at VCIP 2025, represents a significant hardware-software co-design advance. It paves the way for next-generation cameras in smartphones, autonomous vehicles, and robotics that can capture the full luminance range of real-world scenes—from deep shadows to bright highlights—in a single, motion-robust snapshot, eliminating the ghosting and artifacts of current computational HDR methods.