Improving Combined Detection and Classification of TEM Defects via Mask-Conditioned Latent Diffusion Augmentation
Researchers have developed a novel data augmentation technique using a mask-conditioned latent diffusion model to generate synthetic transmission electron microscopy (TEM) images. This method aims to improve the detection and classification of defects in metal alloys, particularly in data-scarce scenarios. By synthesizing realistic images with automatically labeled defect masks, the approach enhances the training of deep learning models, showing modest performance gains in defect analysis. AI
IMPACT Enhances deep learning for microscopy image analysis in data-scarce environments, potentially improving material science research.