Papers › AKDT: Adaptive Kernel Dilation Transformer for Effective Image Denoising

AKDT: Adaptive Kernel Dilation Transformer for Effective Image Denoising

26 Feb 2025VISAPP 2025 2archive 2025-07-28

Alexandru Brateanu, Raul Balmez, Adrian Avram, Ciprian Orhei

Image denoising is a fundamental yet challenging task, especially when dealing with high-resolution images and complex noise patterns. Most existing methods rely on standard Transformer architectures, which often suffer from high computational complexity and limited adaptability to varying noise levels. In this paper, we introduce the Adaptive Kernel Dilation Transformer (AKDT), a novel Transformer-based model that fully harnesses the power of learnable dilation rates within convolutions. AKDT consists of several layers and custom-designed blocks, including our novel Learnable Dilation Rate (LDR) module, which is utilized to construct a Noise Estimator module (NE). At the core of AKDT, the NE is seamlessly integrated within standard Transformer components to form the Noise-Guided Feed-Forward Network (NG-FFN) and Noise-Guided Multi-Headed Self-Attention (NG-MSA). These noise-modulated Transformer components enable the model to achieve unparalleled denoising performance while significantly reducing computational costs. Extensive experiments across multiple image denoising benchmarks demonstrate that AKDT sets a new state-of-the-art, effectively handling both real and synthetic noise. The source code and pre-trained models are publicly available at https://github.com/albrateanu/AKDT.

PaperPDFConference PDFCode

Code

albrateanu/AKDT mentioned in paperpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Color Image DenoisingDenoisingImage Denoising

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Color Image Denoising CBSD68 sigma15 AKDT PSNR 34.64 #1 of 10 Archive leaderboard report
Color Image Denoising CBSD68 sigma25 AKDT PSNR 31.94 #1 of 9 Archive leaderboard report
Color Image Denoising CBSD68 sigma50 AKDT PSNR 28.68 #5 of 18 Archive leaderboard report
Color Image Denoising McMaster sigma15 AKDT PSNR 36.71 #1 of 4 Archive leaderboard report
Color Image Denoising McMaster sigma25 AKDT PSNR 34.21 #1 of 4 Archive leaderboard report
Color Image Denoising McMaster sigma50 AKDT PSNR 30.95 #1 of 7 Archive leaderboard report
Color Image Denoising Urban100 sigma25 AKDT PSNR 33.14 #2 of 6 Archive leaderboard report
Color Image Denoising Urban100 sigma50 AKDT PSNR 29.82 #7 of 9 Archive leaderboard report
Color Image Denoising urban100 sigma15 AKDT Average PSNR 35.63 #1 of 8 Archive leaderboard report
Color Image Denoising urban100 sigma15 AKDT PSNR 35.63 #1 of 8 Archive leaderboard report
Image Denoising SIDD AKDT PSNR (sRGB) 39.70 #13 of 22 Archive leaderboard report
Image Denoising SIDD AKDT SSIM (sRGB) 0.961 #13 of 22 Archive leaderboard report
Image Denoising Urban100 sigma50 AKDT PSNR 29.82 #4 of 4 Archive leaderboard report
Image Denoising urban100 sigma15 AKDT Average PSNR 35.64 #1 of 4 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

Absolute Position EncodingsAdamAttentionBPEDense ConnectionsDropoutLabel SmoothingLayer NormalizationLinear LayerMulti-Head AttentionPosition-Wise Feed-Forward LayerResidual ConnectionSoftmaxTransformer

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections