RIFLE: Removal of Image Flicker-Banding via Latent Diffusion Enhancement

TL;DR

Flicker-banding (FB) degrades photos of emissive displays by temporal aliasing between rolling-shutter capture and display modulation. The paper introduces RIFLE, a one-step diffusion-based restoration framework augmented with a Flicker-banding Prior Estimator (FPE) and a region-focused Masked Loss (ML), plus a comprehensive FB simulation pipeline and a pixel-aligned real-world testing set. The authors demonstrate that RIFLE outperforms recent baselines across multiple metrics on real FB data, with ablations confirming the benefits of ML and FPE. This work provides a solid foundation for FB dataset creation and diffusion-based restoration, with practical impact for improving the readability and quality of screen-captured images.

Abstract

Capturing screens is now routine in our everyday lives. But the photographs of emissive displays are often influenced by the flicker-banding (FB), which is alternating bright%u2013dark stripes that arise from temporal aliasing between a camera's rolling-shutter readout and the display's brightness modulation. Unlike moire degradation, which has been extensively studied, the FB remains underexplored despite its frequent and severe impact on readability and perceived quality. We formulate FB removal as a dedicated restoration task and introduce Removal of Image Flicker-Banding via Latent Diffusion Enhancement, RIFLE, a diffusion-based framework designed to remove FB while preserving fine details. We propose the flicker-banding prior estimator (FPE) that predicts key banding attributes and injects it into the restoration network. Additionally, Masked Loss (ML) is proposed to concentrate supervision on banded regions without sacrificing global fidelity. To overcome data scarcity, we provide a simulation pipeline that synthesizes FB in the luminance domain with stochastic jitter in banding angle, banding spacing, and banding width. Feathered boundaries and sensor noise are also applied for a more realistic simulation. For evaluation, we collect a paired real-world FB dataset with pixel-aligned banding-free references captured via long exposure. Across quantitative metrics and visual comparisons on our real-world dataset, RIFLE consistently outperforms recent image reconstruction baselines from mild to severe flicker-banding. To the best of our knowledge, it is the first work to research the simulation and removal of FB. Our work establishes a great foundation for subsequent research in both the dataset construction and the removal model design. Our dataset and code will be released soon.

Figures (7)

• Figure 1: Overview of RIFLE. The top part presents the datasets we construct, including the simulated dataset (first row) for training and the real-world dataset (second row) for testing. The bottom part shows the flicker-banding removal effect on simulated and real-world images.
• Figure 2: Flicker-banding when filming screens with smartphone cameras. a) Rolling shutter exposure process. b) Typical display brightness modulation (e.g., PWM and scanning refresh). c) Interaction between camera exposure and screen modulation leads to banding artifacts. d) Example banding patterns captured from different display technologies.
• Figure 3: Our flicker-banding simulation pipeline design. a) Stage 1: We generate the general framework based on the basic banding parameters and introduce parameter jitter and feather concatenation for a more realistic transition. b) Stage 2: We overlay the flicker-banding mask on the Y-channel of high-quality (HQ) images and add sensor noise to the reconstructed images.
• Figure 4: Visual comparison between our simulated flicker-banding and real-world flicker-banding. GT indicates the real-world non-banding images, on which our simulation pipeline is conducted. SIM indicates our simulation FB images, while LQ indicates real-world FB images.
• Figure 5: Overview of our model design. a) We train a banding prior estimator (BPE) to predict the banding parameters of low-quality (LQ) inputs. b) We introduce the pretrained BPE to the diffusion structure, resulting in more banding priors for the model. c) We propose a masked loss (ML) to guide the model to focus more on the reconstruction of the image content in the banding area.