Random 2.5D U-net for Fully 3D Segmentation
This addresses the problem of high GPU memory requirements for 3D segmentation in medical imaging or similar fields, offering a more efficient solution.
The paper tackles the computational inefficiency of 3D convolutions for volumetric segmentation by introducing a network that uses 2D convolutions on projections from different directions, avoiding memory issues and outperforming standard approaches with improved resistance to artefacts.
Convolutional neural networks are state-of-the-art for various segmentation tasks. While for 2D images these networks are also computationally efficient, 3D convolutions have huge storage requirements and therefore, end-to-end training is limited by GPU memory and data size. To overcome this issue, we introduce a network structure for volumetric data without 3D convolution layers. The main idea is to include projections from different directions to transform the volumetric data to a sequence of images, where each image contains information of the full data. We then apply 2D convolutions to these projection images and lift them again to volumetric data using a trainable reconstruction algorithm. The proposed architecture can be applied end-to-end to very large data volumes without cropping or sliding-window techniques. For a tested sparse binary segmentation task, it outperforms already known standard approaches and is more resistant to generation of artefacts.