CompareNet: Anatomical Segmentation Network with Deep Non-local Label Fusion

Abstract

Label propagation is a popular technique for anatomical segmentation. In this work, we propose a novel deep framework for label propagation based on non-local label fusion. Our framework, named CompareNet, incorporates subnets for both extracting discriminating features, and learning the similarity measure, which lead to accurate segmentation. We also introduce the voxel-wise classification as an unary potential to the label fusion function, for alleviating the search failure issue of the existing non-local fusion strategies. Moreover, CompareNet is endto-end trainable, and all the parameters are learnt together for the optimal performance. By evaluating CompareNet on two public datasets IB-SRv2 and MICCAI 2012 for brain segmentation, we show it outperforms state-of-the-art methods in accuracy, while being robust to pathologies.

OralViewer: 3D Demonstration of Dental Surgeries for Patient Education with Oral Cavity Reconstruction from a 2D Panoramic X-ray

“…The input 2D panoramic X-ray X ∈ R H ×W (Figure 2(a)), where H and W are image height and width, is fed to a feature extraction subnet (Figure 2(a)) of a 2D encoder-decoder structure for capturing a deep feature map of the same resolution as the input X-ray. Given the feature map, a segmentation subnet (Figure 2(b)) formulates an anatomical segmentation task [25,26] to map it into a categorical mask Y seg ∈ Z H ×W ×K , where K = 32 denotes the maximum number of tooth category. Figure 4a demonstrates the tooth numbering rule we used by following the World Dental Federation Notion [20].…”

TumorCP: A Simple but Effective Object-Level Data Augmentation for Tumor Segmentation

“…We use the implementation of batchgenerators python library, with the alpha range (0, 900) and sigma range (9,13).…”

“…To improve data-efficient learning, several successful approaches have been proposed from different perspectives, such as leveraging unlabeled data for semisupervised self-training [16,1,15] or self-supervised pre-training [19,1,12], distilling priors from data as explicit constraints for model training [10,9], generating new data with the imaging of an anatomy of a different modality [8,11], or utilizing appropriate data augmentation methods to increase data diversity [5,14,2,15]. Some of them are designated for medical images.…”

Generalized Organ Segmentation by Imitating One-shot Reasoning using Anatomical Correlation

“…A large body of literature [1,19,20] exploits anatomical correlation for medical image segmentation within the deep learning framework. The anatomical correlation also serves as the foundation of atlas-based segmentation [12,15], where one or several labeled reference images (i.e., atlases) are non-rigidly registered to a target image based on the anatomical similarity, and the labels of the atlases are propagated to the target image as the segmentation output. Different from these methods, OrganNet has the ability to learn anatomical similarity between images by employing the reasoning process.…”

“…Anatomical similarity have been widely used in medical image segmentation [7,15]. Compared to these methods, our work mainly exploits using anatomical similarity within each one-shot pair of images to perform reasoning.…”