Abstract:The accurate segmentation of brain tumor is significant in clinical practice. Convolutional Neural Network (CNN)-based methods have made great progress in brain tumor segmentation due to powerful local modeling ability. However, brain tumors are frequently pattern-agnostic, i.e. variable in shape, size and location, which can not be effectively matched by traditional CNN-based methods with local and regular receptive fields. To address the above issues, we propose a shape-scale co-awareness network (S2CA-Net) for brain tumor segmentation, which can efficiently learn shape-aware and scale-aware features simultaneously to enhance pattern-agnostic representations. Primarily, three key components are proposed to accomplish the co-awareness of shape and scale. The Local-Global Scale Mixer (LGSM) decouples the extraction of local and global context by adopting the CNN-Former parallel structure, which contributes to obtaining finer hierarchical features. The Multi-level Context Aggregator (MCA) enriches the scale diversity of input patches by modeling global features across multiple receptive fields. The Multi-Scale Attentive Deformable Convolution (MS-ADC) learns the target deformation based on the multiscale inputs, which motivates the network to enforce feature constraints both in terms of scale and shape for optimal feature matching. Overall, LGSM and MCA focus on enhancing the scale-awareness of the network to cope with the size and location variations, while MS-ADC focuses on capturing deformation information for optimal shape matching. Finally, their effective integration prompts the network to perceive variations in shape and scale simultaneously, which can robustly tackle the variations in patterns of brain tumors. The experimental results on BraTS 2019, BraTS 2020, MSD BTS Task and BraTS2023-MEN show that S2CA-Net has superior overall performance in accuracy and efficiency compared to other state-of-the-art methods. Code: https://github.com/jiangyu945/S2CA-Net.

Cervical Spondylotic Myelopathy Segmentation Using Shape-Aware U-net

Improved UNet‐based Magnetic Resonance Imaging Segmentation of Demyelinating Diseases with Small Lesion Regions

Shape-Scale Co-Awareness Network for 3D Brain Tumor Segmentation.

A Robust Segmentation Method Based on Improved U-Net

The Dilated Dense U-net for Spinal Fracture Lesions Segmentation

Deep Learning-Based Auto-Segmentation of Spinal Cord Internal Structure of Diffusion Tensor Imaging in Cervical Spondylotic Myelopathy

Automatic spinal cord segmentation from axial-view MRI slices using CNN with grayscale regularized active contour propagation

Enhancing Spinal MRI Segmentation with an Asymmetric U-Net Architecture

Automated Cervical Tumor Segmentation on MR Images Using Multi-View Feature Attention Network

CSCA U-Net: A channel and space compound attention CNN for medical image segmentation

Automatic Localization And Segmentation Of Vertebral Bodies In 3d Ct Volumes With Deep Learning

Spine Image Segmentation Based on U-Net and Atrous Spatial Pyramid Pooling

DAAM-Net: A dual-encoder U-Net network with adjacent auxiliary module for pituitary tumor and jaw cyst segmentation

AugMS-Net:Augmented Multiscale Network for Small Cervical Tumor Segmentation from MRI Volumes

Accurate Segmentation of Lumbar Spine Magnetic Resonance Imaging by Deep Neural Network with Size-Adaptive Loss.

DSU-net: Dense SegU-net for automatic head-and-neck tumor segmentation in MR images

CM-SegNet: A Deep Learning-Based Automatic Segmentation Approach for Medical Images by Combining Convolution and Multilayer Perceptron

CS2-Net: Deep Learning Segmentation of Curvilinear Structures in Medical Imaging

Automatic Lumbar Spinal CT Image Segmentation With a Dual Densely Connected U-Net

Hierarchical Spatial Pyramid Network for Cervical Precancerous Segmentation by Reconstructing Deep Segmentation Networks.