Saturation in Snapshot Compressive Imaging

Mengyu Zhao, Shirin JalaliarXiv 2025 · DOI · arXiv

Abstract

Snapshot Compressive Imaging (SCI) enables the capture of high-dimensional data using standard 2D sensors by multiplexing multiple frames into a single exposure. However, sensor saturation—where total intensity exceeds the sensor's dynamic range—introduces non-linearities that degrade reconstruction. This paper provides the first theoretical framework to address this issue, deriving a new recovery bound and a design rule favoring sparser optical masks. We also introduce SAPnet, a saturation-aware reconstruction algorithm that utilizes inequality constraints for saturated pixels. SAPnet demonstrates significant performance gains, outperforming existing methods by up to 30 dB in high-saturation scenarios.

Introduction

Snapshot Compressive Imaging (SCI) is an efficient technique for capturing high-speed video or hyperspectral data. By using a coded mask to multiplex frames into a single snapshot, SCI bypasses the bandwidth limitations of traditional sensors. A persistent challenge in practical SCI systems is sensor saturation. When the combined intensity of multiplexed frames exceeds the sensor's full-well capacity, the signal is clipped. This clipping creates a non-linear distortion that standard linear reconstruction algorithms cannot handle, often resulting in "blown-out" regions or artifacts.

Visual impact of varying saturation levels (T/B) on SCI snapshots across different datasets.

Theoretical Framework and Design Principles

The authors model sensor clipping as an element-wise nonlinearity and derive a finite-sample recovery bound. This mathematical foundation reveals that reconstruction error is intrinsically linked to mask density and saturation severity.

A key finding is the "Sparser is Better" rule. While traditional SCI theory suggests mask densities should be around 0.5, this research proves that as saturation increases, the optimal mask density should decrease. In high-brightness environments, a sparser mask (e.g., density of 0.2 or 0.3) blocks more light, preventing the sensor from reaching its saturation threshold and preserving more reconstructible information.

SAPnet: Saturation-Aware Reconstruction

To address saturation algorithmically, the authors developed SAPnet (Saturation-Aware PnP Net). Unlike standard Plug-and-Play (PnP) frameworks that attempt to match reconstructed pixels exactly to clipped measurements, SAPnet treats saturated pixels as inequality constraints. For any pixel $y$ at the saturation threshold $T$, the algorithm enforces that the reconstructed value must be $\ge T$. This "consistency" approach is far more effective than simply rejecting saturated data. SAPnet integrates a pre-trained deep denoiser (FastDVDnet) within this iterative framework, ensuring structural fidelity without requiring retraining for different saturation levels.

Experimental Results

The approach was evaluated on six grayscale video SCI benchmarks: Kobe, Runner, Drop, Traffic, Aerial, and Crash ($256 \times 256$ resolution, $B=8$ frames). These datasets represent varying levels of brightness and motion.

SAPnet achieved massive PSNR gains over standard PnP-FastDVDnet, particularly in highly saturated scenarios ($T/B = 0.25$). For instance, the 'Drop' dataset saw an improvement of +30.017 dB. Qualitatively, SAPnet recovers fine structural details in bright areas that appear as flat blobs in standard reconstructions.

PSNR improvements (dB) of SAPnet over standard PnP-FastDVDnet across different datasets and saturation thresholds (T/B).

Discussion and Conclusion

This work establishes that Bernoulli mask density should generally stay below 0.5, with lower densities being optimal under saturation. The success of SAPnet highlights the power of consistency-based constraints over simple data rejection. While the current implementation relies on a fixed denoiser and Bernoulli masks, future work could extend this theory to optimized "blue noise" masks and end-to-end learned models. This research provides a critical step toward making SCI systems robust for real-world hardware operating under unpredictable lighting conditions.

Summary updated

Abstract

Snapshot compressive imaging (SCI) acquires high-dimensional data cubes, such as videos and hyperspectral images, by optically multiplexing multiple coded frames into a single two-dimensional measurement. While this multiplexing enables high acquisition efficiency, it also increases the risk of sensor saturation: the accumulated intensity may exceed the detector dynamic range, causing clipped measurements that violate the standard linear SCI model. This paper studies SCI reconstruction under such saturated measurements from both theoretical and algorithmic perspectives. We model saturation as an element-wise clipping nonlinearity and derive a finite-sample recovery bound for compression-based SCI. The bound explicitly relates the reconstruction error to the Bernoulli mask density, the compression rate of the signal class, measurement noise, and the expected fraction of saturated measurements. The analysis reveals a principled mask-design rule: under saturation, the optimal Bernoulli mask density remains below one-half and decreases as saturation becomes stronger. Motivated by this result, we optimize mask patterns for saturated acquisition and introduce a saturation-aware plug-and-play reconstruction framework, termed \emph{Saturation-Aware PnP Net} (SAPnet), which enforces consistency with both unsaturated and clipped measurements. Experiments on standard video SCI benchmarks validate the theoretical predictions and show that SAPnet substantially improves reconstruction quality over conventional PnP-based methods, especially in strongly saturated regimes.
BibTeX
@misc{zhao2025saturation,
  author = {Mengyu Zhao and Shirin Jalali},
  title = {{Saturation in Snapshot Compressive Imaging}},
  year = {2025},
  eprint = {2501.11869},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2501.11869},
}
Figure