Generative augmentations for improved cardiac ultrasound segmentation using diffusion models
This work addresses the challenge of generalizing segmentation models in cardiac ultrasound for medical applications, though it is incremental as it builds on existing augmentation methods with diffusion models.
The paper tackled the problem of limited and varied labeled datasets in cardiac ultrasound segmentation by using diffusion models to create generative augmentations, which improved segmentation robustness by over 20 millimeters in Hausdorff distance and enhanced ejection fraction estimation by up to 20% on out-of-distribution cases.
One of the main challenges in current research on segmentation in cardiac ultrasound is the lack of large and varied labeled datasets and the differences in annotation conventions between datasets. This makes it difficult to design robust segmentation models that generalize well to external datasets. This work utilizes diffusion models to create generative augmentations that can significantly improve diversity of the dataset and thus the generalisability of segmentation models without the need for more annotated data. The augmentations are applied in addition to regular augmentations. A visual test survey showed that experts cannot clearly distinguish between real and fully generated images. Using the proposed generative augmentations, segmentation robustness was increased when training on an internal dataset and testing on an external dataset with an improvement of over 20 millimeters in Hausdorff distance. Additionally, the limits of agreement for automatic ejection fraction estimation improved by up to 20% of absolute ejection fraction value on out of distribution cases. These improvements come exclusively from the increased variation of the training data using the generative augmentations, without modifying the underlying machine learning model. The augmentation tool is available as an open source Python library at https://github.com/GillesVanDeVyver/EchoGAINS.