AWDiff: An a trous wavelet diffusion model for lung ultrasound image synthesis

TL;DR

AWDiff is a novel diffusion-based augmentation method for lung ultrasound images that preserves fine details and aligns with clinical labels, improving data quality for machine learning applications.

Contribution

Introduces AWDiff, combining wavelet transforms and semantic conditioning to enhance ultrasound image synthesis while maintaining diagnostic features.

Findings

Lower distortion and higher perceptual quality than existing methods

Preserves fine-scale structures like B-lines and pleural irregularities

Demonstrates improved structural fidelity and clinical diversity

Abstract

Lung ultrasound (LUS) is a safe and portable imaging modality, but the scarcity of data limits the development of machine learning methods for image interpretation and disease monitoring. Existing generative augmentation methods, such as Generative Adversarial Networks (GANs) and diffusion models, often lose subtle diagnostic cues due to resolution reduction, particularly B-lines and pleural irregularities. We propose A trous Wavelet Diffusion (AWDiff), a diffusion based augmentation framework that integrates the a trous wavelet transform to preserve fine-scale structures while avoiding destructive downsampling. In addition, semantic conditioning with BioMedCLIP, a vision language foundation model trained on large scale biomedical corpora, enforces alignment with clinically meaningful labels. On a LUS dataset, AWDiff achieved lower distortion and higher perceptual quality compared to existing methods, demonstrating both structural fidelity and clinical diversity.