Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Edge-preserving Image Denoising via Multi-scale Adaptive Statistical Independence Testing

Ruyu Yan, Da-Qing Zhang

TL;DR

This work tackles robust, edge-preserving image denoising in noisy and multi-scale contexts by introducing EDD-MAIT, which fuses channel attention with gradient-driven, multi-scale adaptive statistical independence testing. The method combines gradient-based preliminary edge cues, adaptive windowing guided by local complexity, and statistical tests (Fisher's exact test or chi-square test) with Otsu-based dual thresholds to distinguish true edges from noise. Key contributions include a gradient-informed window adaptation strategy, a probability-based edge confirmation via independence testing, and an adaptive denoising mechanism that achieves competitive or superior performance on BSDS500 and BIPED with favorable runtime. The approach demonstrates robustness to Gaussian noise and offers practical gains for edge-aware applications in computer vision where efficiency and accuracy are critical.

Abstract

Edge detection is crucial in image processing, but existing methods often produce overly detailed edge maps, affecting clarity. Fixed-window statistical testing faces issues like scale mismatch and computational redundancy. To address these, we propose a novel Multi-scale Adaptive Independence Testing-based Edge Detection and Denoising (EDD-MAIT), a Multi-scale Adaptive Statistical Testing-based edge detection and denoising method that integrates a channel attention mechanism with independence testing. A gradient-driven adaptive window strategy adjusts window sizes dynamically, improving detail preservation and noise suppression. EDD-MAIT achieves better robustness, accuracy, and efficiency, outperforming traditional and learning-based methods on BSDS500 and BIPED datasets, with improvements in F-score, MSE, PSNR, and reduced runtime. It also shows robustness against Gaussian noise, generating accurate and clean edge maps in noisy environments.

Edge-preserving Image Denoising via Multi-scale Adaptive Statistical Independence Testing

TL;DR

This work tackles robust, edge-preserving image denoising in noisy and multi-scale contexts by introducing EDD-MAIT, which fuses channel attention with gradient-driven, multi-scale adaptive statistical independence testing. The method combines gradient-based preliminary edge cues, adaptive windowing guided by local complexity, and statistical tests (Fisher's exact test or chi-square test) with Otsu-based dual thresholds to distinguish true edges from noise. Key contributions include a gradient-informed window adaptation strategy, a probability-based edge confirmation via independence testing, and an adaptive denoising mechanism that achieves competitive or superior performance on BSDS500 and BIPED with favorable runtime. The approach demonstrates robustness to Gaussian noise and offers practical gains for edge-aware applications in computer vision where efficiency and accuracy are critical.

Abstract

Edge detection is crucial in image processing, but existing methods often produce overly detailed edge maps, affecting clarity. Fixed-window statistical testing faces issues like scale mismatch and computational redundancy. To address these, we propose a novel Multi-scale Adaptive Independence Testing-based Edge Detection and Denoising (EDD-MAIT), a Multi-scale Adaptive Statistical Testing-based edge detection and denoising method that integrates a channel attention mechanism with independence testing. A gradient-driven adaptive window strategy adjusts window sizes dynamically, improving detail preservation and noise suppression. EDD-MAIT achieves better robustness, accuracy, and efficiency, outperforming traditional and learning-based methods on BSDS500 and BIPED datasets, with improvements in F-score, MSE, PSNR, and reduced runtime. It also shows robustness against Gaussian noise, generating accurate and clean edge maps in noisy environments.
Paper Structure (23 sections, 19 equations, 10 figures, 6 tables, 1 algorithm)

This paper contains 23 sections, 19 equations, 10 figures, 6 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related Work
  3. Traditional edge detection
  4. Edge detection based on deep learning
  5. Method
  6. Channel Attention Model
  7. Preliminary Edge Detection
  8. Calculate gradients and partition membership degrees
  9. Median filtering coarsening and morphological smoothing of images
  10. Adaptive edge detection method based on independence testing
  11. Calculate the complexity of local regions
  12. Edge Detection Method Based on Independence Test
  13. Otsu threshold segmentation
  14. experiment and result analysis
  15. Experiment
  16. ...and 8 more sections

Figures (10)

  • Figure 1: Illustration of Feature Extraction Using Attention Mechanism.
  • Figure 2: Overview of the proposed EDD-MAIT edge detection framework.
  • Figure 3: Extract continuous edge curves and randomly generated discrete points that can be regarded as noise points from the image.
  • Figure 4: Comparison result diagram of different window sizes
  • Figure 5: Running time and F-measure corresponding to different overlap rates
  • ...and 5 more figures