Efficient 3D Brain Tumor Segmentation with Axial-Coronal-Sagittal Embedding

2025-05-31

Pacific-Rim Symposium on Image and Video Technology

AI Summary

To address the high computational cost and low pretraining weight utilization of nnU-Net in 3D brain tumor segmentation, this work proposes an efficient, lightweight framework. First, we introduce a novel tri-planar (axial/coronal/sagittal) embedded convolutional architecture to enhance spatial contextual modeling. Second, we design two 2D→3D pretraining weight transfer strategies: one leveraging ImageNet-pretrained generic features, and another exploiting glioma grading task-specific encoder representations. Third, we formulate a joint classification-segmentation learning framework to improve robustness for small tumor subregions. Experiments on the BraTS dataset demonstrate that our method reduces training time by 40% and decreases trainable parameters by 32%, while achieving single-model performance comparable to—or even surpassing—that of conventional cross-validation ensemble models. This advancement significantly improves both efficiency and clinical practicality in medical image segmentation.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Ensemble Methods
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• Web Mining and Content Analysis: Large pretrained models with web data
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

In this paper, we address the crucial task of brain tumor segmentation in medical imaging and propose innovative approaches to enhance its performance. The current state-of-the-art nnU-Net has shown promising results but suffers from extensive training requirements and underutilization of pre-trained weights. To overcome these limitations, we integrate Axial-Coronal-Sagittal convolutions and pre-trained weights from ImageNet into the nnU-Net framework, resulting in reduced training epochs, reduced trainable parameters, and improved efficiency. Two strategies for transferring 2D pre-trained weights to the 3D domain are presented, ensuring the preservation of learned relationships and feature representations critical for effective information propagation. Furthermore, we explore a joint classification and segmentation model that leverages pre-trained encoders from a brain glioma grade classification proxy task, leading to enhanced segmentation performance, especially for challenging tumor labels. Experimental results demonstrate that our proposed methods in the fast training settings achieve comparable or even outperform the ensemble of cross-validation models, a common practice in the brain tumor segmentation literature.

Problem

• Improve brain tumor segmentation efficiency using multi-planar convolutions
• Reduce training demands by integrating pre-trained ImageNet weights
• Enhance segmentation via joint classification-segmentation with glioma grade encoders

Innovation

• Integrates Axial-Coronal-Sagittal convolutions into nnU-Net
• Transfers 2D pre-trained weights to 3D domain
• Uses joint classification-segmentation with pre-trained encoders