Defect Image Sample Generation With Diffusion Prior for Steel Surface Defect Recognition

TL;DR

This paper introduces StableSDG, a novel method leveraging diffusion models to generate high-quality steel surface defect images, addressing data scarcity and improving defect recognition accuracy.

Contribution

The paper proposes StableSDG, which adapts diffusion models for steel defect image generation through distribution alignment and image-oriented generation, enhancing sample quality.

Findings

Achieved state-of-the-art defect image generation quality.

Improved defect recognition model performance.

Validated effectiveness on steel surface defect dataset.

Abstract

The task of steel surface defect recognition is an industrial problem with great industry values. The data insufficiency is the major challenge in training a robust defect recognition network. Existing methods have investigated to enlarge the dataset by generating samples with generative models. However, their generation quality is still limited by the insufficiency of defect image samples. To this end, we propose Stable Surface Defect Generation (StableSDG), which transfers the vast generation distribution embedded in Stable Diffusion model for steel surface defect image generation. To tackle with the distinctive distribution gap between steel surface images and generated images of the diffusion model, we propose two processes. First, we align the distribution by adapting parameters of the diffusion model, adopted both in the token embedding space and network parameter space. Besides, in the generation process, we propose image-oriented generation rather than from pure Gaussian noises. We conduct extensive experiments on steel surface defect dataset, demonstrating state-of-the-art performance on generating high-quality samples and training recognition models, and both designed processes are significant for the performance.