Aug 2026· IEEE Transactions on Pattern Analysis and Machine Intelligence· Vol PP· 0 citations· 99 references
Computer ScienceMedicine
TL;DR
This paper proposes BinRVR, a binarized RAW video restoration framework that reduces computation and parameters by approximately 96% while incurring only about 4% performance degradation, and develops a Distribution-Aware Binarized Convolution (DAB-Conv) that leverages the statistics of full-precision activations to mitigate quantization errors.
Abstract
RAW video restoration is fundamental to high-quality low-level perception and serves as the basis for a wide range of downstream vision applications. While binary neural networks (BNNs) enable efficient lightweight deployment for image enhancement, their deficiencies in modeling temporal coherence and activation value distributions hinder their effectiveness when applied to video scenarios. In this paper, we propose BinRVR, a binarized RAW video restoration framework that reduces computation and parameters by approximately 96% while incurring only about 4% performance degradation. Specifically, we present a Binarized Information Interaction Module (BIIM) to jointly model spatial and temporal information in an efficient and unified manner. Moreover, we develop a Distribution-Aware Binarized Convolution (DAB-Conv) that leverages the statistics of full-precision activations to mitigate quantization errors. The proposed framework further supports multi-bit quantization, enabling flexible accuracy-efficiency trade-offs across different hardware constraints. Extensive experiments demonstrate that our BinRVR achieves competitive performance compared with state-of-the-art binarized methods on RAW video restoration tasks, including low-light enhancement, denoising, deblurring, and super-resolution. We further explore the potential of our method on downstream video applications, including object detection and monocular depth estimation.
High-Compression videos suffer from severe distortions, among which degradation in person regions has the greatest impact on viewers’ immersive experience. Existing quality enhancement techniques usually focus on overall image denoising or super-resolution, often overlooking the crucial recovery of fine structures in these essential person regions. To address these challenges, the research introduces a novel framework titled Person Region Restoration Driven by Perceptual Fidelity (PRRDPF), which combines long-range dependency features with perceptual structure loss for enhanced generative restoration. Specifically, first, the research constructs a high-fidelity distorted person-region dataset via a closed-loop degradation pipeline, addressing the lack of paired datasets. Secondly, a Temporal Gated Fusion (TGF) block is designed to use gated convolutions for selectively recovering high-frequency features while capturing local and global dependencies. Finally, a Structural Similarity Index Measure (SSIM)-based dynamic weighted adversarial loss is proposed to prioritize the restoration of visual texture details. Experimental results validate that PRRDPF significantly outperforms the best models in Peak Signal-to-Noise Ratio (PSNR), SSIM, and Learned Perceptual Image Patch Similarity (LPIPS), effectively mitigating artifacts and enhancing clarity in person visuals. This framework presents a promising approach for intelligent video coding integrated with generative artificial intelligence and holds significant potential for practical applications.
Linyun Liu, Li Yu, Jiaxin Zeng et al.· IEEE Signal Processing Lette...· 0 citations
DPCA-Net establishes a superior Pareto front between restoration fidelity and efficiency by maintaining real-time inference speeds with a minimal footprint of 15.00 GFLOPs, offering a compelling practical solution for real-time all-in-one image restoration.
Despite the rapid progress of restoration-aware compression, existing studies still face challenges in low-bitrate restoration scenarios. Real-world image acquisition is inevitably affected by degradations, such as noise, blur, and adverse weather, while subsequent compression further reduces perceptual quality. These coupled degradations make it difficult to recover high-quality images from compact bitstreams. To address this issue, we propose a Decoupled Energy-Complementary method for low-bitrate perceptual restoration-aware compression (DEC). DEC jointly considers acquisition distortions and compression effects within a unified restoration framework. It adopts a decoupled dual-branch architecture with a MainInfo Branch and a SideInfo Branch. The MainInfo Branch preserves structural information for reliable reconstruction, whereas the SideInfo Branch compactly encodes boundary and texture cues with only a few additional bits. These two branches form a compact but informative representation that balances compression efficiency and restoration quality. Extensive experiments show that the SideInfo Branch enhances local restoration details across bitrates, while the MainInfo Branch maintains structural fidelity. Overall, DEC achieves an average 35.49% BD-rate reduction with consistent LPIPS gains over state-of-the-art restoration-aware compression methods, confirming its efficiency and visual quality.
Yuan Xue, Qi Zhang, Shiqi Wang et al.· IEEE Transactions on Image P...· 0 citations
Image dehazing aims to generate the haze-free images from the hazy observation images. While recent deep learning approaches achieve impressive restoration quality, they suffer from excessive computational complexity and model size, hindering practical applications for real-world deployment on resource-constrained edge devices. To address the limitation, lightweight models are proposed to this end but compromise on dehazing performance. To bridge this gap, we propose SLRNet, Super Lightweight Residual Network, a high efficient-yet-effective end-to-end dehazing architecture. SLRNet integrates a novel Adaptive Feature Unit that automatically adjusts channel-wise features through a lightweight gating mechanism, coupled with compact residual blocks to preserve critical structural information. Unlike standard channel attention mechanisms that discard spatial information, our AFU employs an asymmetric split strategy to simultaneously preserve local texture details and capture global haze density. Our design emphasizes minimal parameter count and low latency without sacrificing perceptual quality. Experiments are carried out across standard benchmarks, showing that our proposed SLRNet demonstrates remarkable performance by achieving state-of-the-art efficiency-accuracy trade-offs compared to existing works, while maintaining robust generalization to real-world haze despite the synthetic-to-real domain gap. The codes are released in https://anonymous.4open.science/r/SLRNet.
Guanheng Qu· Poster Volume 0007 The 2026...· 0 citations
The confidence-guided hybrid network (CGHNet) is proposed, a parallel three-branch framework that jointly performs frequency-decoupled local restoration, global context modeling, and pixel-wise degradation prior estimation and its key component is a confidence-guided feature purification mechanism.
Xiaohui Kou, Yang Yan, Qiuyan Wang et al.· Journal of Supercomputing· 0 citations
This study evaluates the viability of deep learningbased super-resolution (SR) for enhancing real-time video streaming. We compare a traditional HEVC-compressed streaming pipeline against an AI-assisted framework that transmits lowresolution video and reconstructs high-resolution output at the client using the Efficient Sub-Pixel Convolutional Neural Network (ESPCN) algorithm. The evaluation spans multiple resolutions (1080p to 4K), motion types, and processing methodologies (subprocess vs. frame-by-frame). Experimental results indicate that while AI-assisted streaming significantly reduces transmission bandwidth, it introduces substantial computational overhead, with completion times increasing by over 300% compared to traditional methods. Furthermore, quality assessments using PSNR and VMAF metrics reveal that most AI-assisted streams failed to meet the target quality range (30–50 dB PSNR), with high-resolution 4K streams experiencing up to 24.58% frame loss during real-time inference. These findings highlight critical performance barriers in utilizing current SR algorithms for standard consumer-grade real-time video delivery.
Emma Hubbell, Hao Wu, Yulei Pang· 2026 International Conferenc...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
