Anti-Aliasing Snapshot HDR Imaging Using Non-Regular Sensing
Teresa Stürzenhofäcker, Moritz Klimm, Jürgen Seiler, André Kaup · 2026
Read Anti-Aliasing Snapshot HDR Imaging Using Non-Regular Sensing by Teresa Stürzenhofäcker, Moritz Klimm, Jürgen Seiler, André Kaup (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-05133.txt.
Introduction
Teresa Stürzenhofäcker and team present an HDR imaging sensor based on spatially varying apertures, mixing two differently sized prototype pixels. Other light joining areas stretch dynamic range toward the lower end versus a standard high detail sensor. Non-regular pixel arrangement mitigates aliasing and overcomes detail loss from the larger prototype pixel's lifted joining area. Fourier-domain rebuild, where natural images are sparse, recovers the image with high detail. Runs suggest the layout acquires high dynamic range images free from aliasing artefacts.
Catalog record
| CREATOR | Teresa Stürzenhofäcker, Moritz Klimm, Jürgen Seiler, André Kaup |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 25237 |
| LICENSE | http://creativecommons.org/licenses/by-sa/4.0/ |
| RIGHTS | CC BY SA |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.05133
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Anti-Aliasing Snapshot HDR Imaging Using Non-Regular Sensing Teresa Stürzenhofäcker, Moritz Klimm, Jürgen Seiler, André Kaup Abstract Snapshot HDR imaging is essential to capture the full dynamic range of a scene in a single exposure, making it essential for video and dynamic environments where motion prevents the use of multi-exposure techniques or complex hardware set-ups. This work presents a snapshot HDR imaging sensor that is based on spatially varying apertures, implemented by combining two differently sized prototype pixels. The different light integration areas physically extend the dynamic range towards the lower end, compared to a standard high resolution sensor. A non-regular pixel arrangement is suggested, to mitigate aliasing and overcome a loss in spatial resolution that is associated with increased light integration area of the larger prototype pixel. Subsequent reconstruction in the Fourier domain, where natural images can be sparsely represented allows to recover the image with high detail. The image acquisition approach with the proposed non-regular HDR sensor is simulated and analysed with special emphasis on the spatial resolution. The results suggest the snapshot HDR sensor layout to be an effective way to acquire images with high dynamic range and free from aliasing artefacts. I Introduction High dynamic range (HDR) imaging is essential for capturing scenes with significant luminance variations. Even though modern high-end cameras have improved their dynamic range over the past few years, it is still not on a par with the human visual system [ 7 ] . A straight forward approach to increase the dynamic range relies on merging multiple exposures of the same scene which works very well for static scenes but fails to capture motion [ 3 ] . For moving objects, beam splitters can be used to capture synchronised exposures with multiple sensors [ 1 ] . However, this requires a more expensive hardware set-up and higher complexity. Alternatively, Dual-Conversion Gain (DCG) applies two different conversion gains to the same photodiode and thereby allows to capture both bright and dark areas [ 2 ] . Yet, this snapshot approach does not change the number of collected photons and therefore does not extend the physical dynamic range [ 2 ] . Another snapshot approach uses neutral density (ND) filters on individual pixels to physically extend the dynamic range by controlling the number of collected photons [ 13 ] . While effective for capturing highlights, this method cannot improve dark scene acquisition, as modern cameras already approach photometric limits [ 15 ] . The only way a sensor can reduce the impact of noise in darker scenes is by using pixels with larger light integration areas. Fujifilm introduced the SuperCCD SR sensor which is based on hexagonal pixels of different sizes [ 18 ] . The large prototype pixel captures dark areas accurately, while the small pixel collects fewer photons and is thus better suited for bright scene regions. While the SuperCCD SR sensor increases the dynamic range by 400% it has never dominated the camera market due to major drawbacks regarding the spatial resolution, accompanied by increased manufacturing complexity due to the hexagonal pixel shapes [ 9 ] . The loss in spatial resolution is a result of the larger pixel size, which in turn decreases the sampling distance and introduces aliasing. Our work is improving on this by proposing a snapshot HDR imaging sensor layout that benefits from differently sized pixels while mitigating aliasing and preserving high-resolution detail, by introducing non-regular pixel arrangements. Consequently, the proposed sensor reduces information loss during image acquisition which usually origins from to exceeded exposures and aliasing. Thereby high image signal fidelity is provided for any further classical or ML-based image processing steps. The novelty of this work is posed by a HDR snapshot sensor design benefiting from • two differently sized prototype pixels that extend the sensors dynamic range physically by 300% • exploiting a non-regular arrangement of the prototype pixels to allow for an aliasing-free image acquisition. The paper is organised as follows: first an overview on the image acquisition pipeline is given in Section II . In Section III , the proposed HDR sensor is presented with its ability to increase the dynamic range and avoid aliasing artefacts. Finally, the simulations and results are provided in Section IV , followed by a summary in Section V . II Image Acquisition model To capture an image, the scene radiance \Phi has to be converted into digital pixel values through physical and electronic integrations and processing [ 5 ] . Radiance from a scene, \Phi_{\text{scene}} , is focused onto the sensor by the lens, which integrates incoming light over a solid angle \Omega defined by the aperture. The shutter controls the exposure time, integrating irradiance over time. A measure of the energy at each pixel can be obtained by a spatial integration over the pixel area. In the final step, the image signal is amplified, quantized, digitized, and mapped to pixel values using a non-linear camera response function (CRF) [ 5 ] . The acquisition can be simplified as, \displaystyle I(x_{0},y_{0})\propto\iint_{\mathcal{R}}\int_{\tau_{\text{exp}}}\int_{\Omega} \displaystyle\Phi_{\text{scene}}(x,y,t,\Omega)\,\text{d}\Omega\,\text{d}t\,\text{d}x\,\text{d}y (1) \displaystyle+ \displaystyle N_{\text{dynamic}}(x,y)\, where \mathcal{R} defines the region of a pixel and \tau_{\text{exp}} denotes the time of exposure. In reality, multiple noise sources, combined in N_{\text{dynamic}}(x,y) degrade the image quality during the acquisition process. The most quality degrading noise source is usually shot noise which arises from the statistical variation in the number of photons hitting the sensor [ 16 ] . The shot noise behaviour can be modelled by a poisson distribution which is especially dominant in low-light situations and short exposures when only few photons arrive at the sensor. Other noise sources like dark current, read-out noise and thermal noise are approximated by adding gaussian noise to the acquisition model. Further, the sensor saturation clips the measured intensity I(x_{0},y_{0}) according to, I_{\text{saturated}}(x_{0},y_{0})=\text{min}\left(I_{\text{max}},I(x_{0},y_{0})\right). (2) The maximum value I_{\text{max}} and the amount of present noise limit the dynamic range of the acquired image [ 9 ] . III Proposed High Dynamic Range Sensor Layout Inspired by Fujifilm’s Super CCD [ 18 ] this work also proposes a snapshot HDR sensor design with two prototype pixels of different sizes [ 18 ] . However, unlike the Fuji sensor, the proposed sensor layout is able to acquire images aliasing-free and thereby maintain a high spatial resolution. For this purpose non-regular sampling is suggested, which proofed to capture dominant scene frequencies, even with a reduced number of photometric measurements [ 17 ] . The considered sensor layout thereby combines the two imaging approaches and enables snapshot HDR imaging at a high spatial resolution. To evaluate and demonstrate the impact of non-regular pixel arrangements on the resolution, the proposed snapshot HDR sensor, shown in. 1 a, is compared against a regularly arranged sensor with equivalent prototype pixels, illustrated in Fig. 1 b. (a) Non-Regular Layout (b) Regular Layout Fig. 1 : The proposed, sensor Layout in (a) with non-regular orientation of small pixel (green) and large pixel (red). In contrast, the regular layout in (b) with constant prototype pixel orientation. III-A HDR Imaging with Spatially Varying Pixel Sizes The two prototype pixels interleave perfectly without any gaps or overlaps and achieve a 100% fill factor of the image sensor, which is required for maximum light sensitivity. Due to