Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

ExDD: Explicit Dual Distribution Learning for Surface Defect Detection via Diffusion Synthesis

Muhammad Aqeel, Federico Leonardi, Francesco Setti

TL;DR

The proposed ExDD (Explicit Dual Distribution) approach leverages parallel memory banks that capture the distinct statistical properties of both normality and anomalous patterns, addressing the fundamental flaw of uniform outlier assumptions.

Abstract

Industrial defect detection systems face critical limitations when confined to one-class anomaly detection paradigms, which assume uniform outlier distributions and struggle with data scarcity in real-world manufacturing environments. We present ExDD (Explicit Dual Distribution), a novel framework that transcends these limitations by explicitly modeling dual feature distributions. Our approach leverages parallel memory banks that capture the distinct statistical properties of both normality and anomalous patterns, addressing the fundamental flaw of uniform outlier assumptions. To overcome data scarcity, we employ latent diffusion models with domain-specific textual conditioning, generating in-distribution synthetic defects that preserve industrial context. Our neighborhood-aware ratio scoring mechanism elegantly fuses complementary distance metrics, amplifying signals in regions exhibiting both deviation from normality and similarity to known defect patterns. Experimental validation on KSDD2 demonstrates superior performance (94.2% I-AUROC, 97.7% P-AUROC), with optimal augmentation at 100 synthetic samples. https://github.com/aqeeelmirza/ExDD-Defect-Detection

ExDD: Explicit Dual Distribution Learning for Surface Defect Detection via Diffusion Synthesis

TL;DR

The proposed ExDD (Explicit Dual Distribution) approach leverages parallel memory banks that capture the distinct statistical properties of both normality and anomalous patterns, addressing the fundamental flaw of uniform outlier assumptions.

Abstract

Industrial defect detection systems face critical limitations when confined to one-class anomaly detection paradigms, which assume uniform outlier distributions and struggle with data scarcity in real-world manufacturing environments. We present ExDD (Explicit Dual Distribution), a novel framework that transcends these limitations by explicitly modeling dual feature distributions. Our approach leverages parallel memory banks that capture the distinct statistical properties of both normality and anomalous patterns, addressing the fundamental flaw of uniform outlier assumptions. To overcome data scarcity, we employ latent diffusion models with domain-specific textual conditioning, generating in-distribution synthetic defects that preserve industrial context. Our neighborhood-aware ratio scoring mechanism elegantly fuses complementary distance metrics, amplifying signals in regions exhibiting both deviation from normality and similarity to known defect patterns. Experimental validation on KSDD2 demonstrates superior performance (94.2% I-AUROC, 97.7% P-AUROC), with optimal augmentation at 100 synthetic samples. https://github.com/aqeeelmirza/ExDD-Defect-Detection

Paper Structure

This paper contains 23 sections, 17 equations, 2 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. ExDD Framework
  4. Problem Formulation
  5. Dual Memory Bank Architecture
  6. Locally Aware Patch Features
  7. Negative and Positive Memory Banks
  8. Dimensionality Reduction and Coreset Subsampling
  9. Diffusion-based Anomaly Generation
  10. Theoretical Foundation
  11. Synthetic Anomaly Generation Pipeline
  12. Anomaly Detection with ExDD
  13. Distance Computation
  14. Neighborhood-Aware Weighting
  15. Ratio Scoring
  16. ...and 8 more sections

Figures (2)

  • Figure 1: Overview of the ExDD framework, illustrating the training process with pretrained encoder and patch feature extraction, synthetic anomaly generation using diffusion models with prompt guidance, testing workflow, and the dual memory bank architecture with ratio-based anomaly scoring mechanism.
  • Figure 2: Qualitative comparison of anomaly localization results on the KSDD2 test set.