MEDDAP: Medical Dataset Enhancement via Diversified Augmentation Pipeline
Yasamin Medghalchi, Niloufar Zakariaei, Arman Rahmim, Ilker Hacihaliloglu
TL;DR
This work tackles limited annotated ultrasound data and acquisition variability by introducing MEDDAP, a diffusion-based augmentation pipeline guided by Ultrasound Low-Rank Adaptation (USLoRA). By fine-tuning cross-attention in Stable Diffusion with a low-rank update and enriching prompts with adjectives, MEDDAP generates diverse, informative labeled ultrasound samples that improve downstream breast cancer classification and robustness to unseen data. The method demonstrates improved accuracy and reliability across CNN architectures and datasets, supported by quantitative metrics (FID) and qualitative heat-map analyses, while reducing computational demands through targeted fine-tuning. Overall, MEDDAP offers a practical, resource-efficient path for expanding medical imaging datasets and enhancing model generalization in ultrasound-based diagnostics, with code publicly available.
Abstract
The effectiveness of Deep Neural Networks (DNNs) heavily relies on the abundance and accuracy of available training data. However, collecting and annotating data on a large scale is often both costly and time-intensive, particularly in medical cases where practitioners are already occupied with their duties. Moreover, ensuring that the model remains robust across various scenarios of image capture is crucial in medical domains, especially when dealing with ultrasound images that vary based on the settings of different devices and the manual operation of the transducer. To address this challenge, we introduce a novel pipeline called MEDDAP, which leverages Stable Diffusion (SD) models to augment existing small datasets by automatically generating new informative labeled samples. Pretrained checkpoints for SD are typically based on natural images, and training them for medical images requires significant GPU resources due to their heavy parameters. To overcome this challenge, we introduce USLoRA (Ultrasound Low-Rank Adaptation), a novel fine-tuning method tailored specifically for ultrasound applications. USLoRA allows for selective fine-tuning of weights within SD, requiring fewer than 0.1\% of parameters compared to fully fine-tuning only the UNet portion of SD. To enhance dataset diversity, we incorporate different adjectives into the generation process prompts, thereby desensitizing the classifiers to intensity changes across different images. This approach is inspired by clinicians' decision-making processes regarding breast tumors, where tumor shape often plays a more crucial role than intensity. In conclusion, our pipeline not only outperforms classifiers trained on the original dataset but also demonstrates superior performance when encountering unseen datasets. The source code is available at https://github.com/yasamin-med/MEDDAP.