Deep Q-Network-Driven Catheter Segmentation in 3D US by Hybrid Constrained Semi-Supervised Learning and Dual-UNet

TL;DR

This paper introduces a semi-supervised deep learning approach for 3D ultrasound catheter segmentation that reduces annotation needs and improves accuracy by combining deep Q-learning for localization with a Dual-UNet for segmentation.

Contribution

It presents a novel hybrid constrained semi-supervised learning framework integrating deep Q-learning and Dual-UNet for efficient catheter segmentation with limited labeled data.

Findings

Achieves higher segmentation performance than state-of-the-art semi-supervised methods.

Effectively leverages large-scale unlabeled images for training.

Reduces annotation effort while maintaining high accuracy.

Abstract

Catheter segmentation in 3D ultrasound is important for computer-assisted cardiac intervention. However, a large amount of labeled images are required to train a successful deep convolutional neural network (CNN) to segment the catheter, which is expensive and time-consuming. In this paper, we propose a novel catheter segmentation approach, which requests fewer annotations than the supervised learning method, but nevertheless achieves better performance. Our scheme considers a deep Q learning as the pre-localization step, which avoids voxel-level annotation and which can efficiently localize the target catheter. With the detected catheter, patch-based Dual-UNet is applied to segment the catheter in 3D volumetric data. To train the Dual-UNet with limited labeled images and leverage information of unlabeled images, we propose a novel semi-supervised scheme, which exploits unlabeled images based on hybrid constraints from predictions. Experiments show the proposed scheme achieves a higher performance than state-of-the-art semi-supervised methods, while it demonstrates that our method is able to learn from large-scale unlabeled images.