Copy-Paste Image Augmentation with Poisson Image Editing for Ultrasound Instance Segmentation Learning

TL;DR

This paper introduces a novel copy-paste augmentation method using Poisson image editing for ultrasound images, enhancing training data diversity and improving segmentation performance.

Contribution

It presents a new augmentation technique that creates realistic ultrasound images by seamlessly blending pasted regions, which was not previously explored for this application.

Findings

Improved segmentation accuracy with the proposed augmentation.

More stable training process observed.

Higher objective metrics achieved during training.

Abstract

Deep learning has shown great success in high-level image analysis problems; yet its efficacy relies on the quality and diversity of the training data. In this work, we introduce a copypaste image augmentation for ultrasound images. The Poisson image editing technique is used to generate realistic and seamless boundary transitions around the pasted image. Results showed that the proposed image augmentation technique improves training performance in terms of higher objective metrics and more stable training results.