Efficient multi-scale 3D CNN with fully connected CRF for accurate brain lesion segmentation

TLDR

The architecture results from an in-depth analysis of limitations in current networks for similar applications. The study proposes a dual‑pathway, 11‑layer 3D CNN for brain lesion segmentation and introduces an efficient dense‑training scheme that processes adjacent patches in a single pass while addressing class imbalance. The method uses a dual‑pathway 11‑layer 3D CNN that processes multi‑scale inputs, applies a 3D fully connected CRF for false‑positive removal, and is evaluated on traumatic brain injury, brain tumor, and ischemic stroke MRI datasets. The approach outperforms state‑of‑the‑art on BRATS 2015 and ISLES 2015, is computationally efficient for clinical use, and its code is publicly available.

Abstract

We propose a dual pathway, 11-layers deep, three-dimensional Convolutional Neural Network for the challenging task of brain lesion segmentation. The devised architecture is the result of an in-depth analysis of the limitations of current networks proposed for similar applications. To overcome the computational burden of processing 3D medical scans, we have devised an efficient and effective dense training scheme which joins the processing of adjacent image patches into one pass through the network while automatically adapting to the inherent class imbalance present in the data. Further, we analyze the development of deeper, thus more discriminative 3D CNNs. In order to incorporate both local and larger contextual information, we employ a dual pathway architecture that processes the input images at multiple scales simultaneously. For post-processing of the network's soft segmentation, we use a 3D fully connected Conditional Random Field which effectively removes false positives. Our pipeline is extensively evaluated on three challenging tasks of lesion segmentation in multi-channel MRI patient data with traumatic brain injuries, brain tumours, and ischemic stroke. We improve on the state-of-the-art for all three applications, with top ranking performance on the public benchmarks BRATS 2015 and ISLES 2015. Our method is computationally efficient, which allows its adoption in a variety of research and clinical settings. The source code of our implementation is made publicly available.