Abstract:The precise segmentation of bladder tumors from MRI is essential for bladder cancer diagnosis and personalized therapy selection. Limited by the properties of tumor morphology, achieving precise segmentation from MRI images remains challenging. In recent years, deep convolutional neural networks have provided a promising solution for bladder tumor segmentation from MRI. However, deep-learning-based methods still face two weakness: (1) multi-scale feature extraction and utilization are inadequate, being limited by the learning approach. (2) The establishment of explicit long-distance dependence is difficult due to the limited receptive field of convolution kernels. These limitations raise challenges in the learning of global semantic information, which is critical for bladder cancer segmentation. To tackle the problem, a newly auxiliary segmentation algorithm integrating a multi-scale encoder and decoder with a transformer is proposed, which is called MSEDTNet. Specifically, the designed encoder with multi-scale pyramidal convolution (MSPC) is utilized to generate compact feature maps which capture the richly detailed local features of the image. Furthermore, the transformer bottleneck is then leveraged to model the long-distance dependency between high-level tumor semantics from a global space. Finally, a decoder with a spatial context fusion module (SCFM) is adopted to fuse the context information and gradually produce high-resolution segmentation results. The experimental results of T2-weighted MRI scans from 86 patients show that MSEDTNet achieves an overall Jaccard index of 83.46%, a Dice similarity coefficient of 92.35%, and a complexity less than that of other, similar models. This suggests that the method proposed in this article can be used as an efficient tool for clinical bladder cancer segmentation.

Dense Encoder-Decoder Network Based On Two-Level Context Enhanced Residual Attention Mechanism For Segmentation Of Breast Tumors In Magnetic Resonance Image

Deep Learning-Based Automatic Segmentation for Size and Volumetric Measurement of Breast Cancer on Magnetic Resonance Imaging.

Joint Dense Residual and Recurrent Attention Network for DCE-MRI Breast Tumor Segmentation

DCNet: Densely Connected Deep Convolutional Encoder–Decoder Network for Nasopharyngeal Carcinoma Segmentation

Encoder–Decoder Network with Depthwise Atrous Spatial Pyramid Pooling for Automatic Brain Tumor Segmentation

MLRD-Net: 3D multiscale local cross-channel residual denoising network for MRI-based brain tumor segmentation

Multi-scale Strategy Based 3D Dual-Encoder Brain Tumor Segmentation Network with Attention Mechanism.

MSEDTNet: Multi-Scale Encoder and Decoder with Transformer for Bladder Tumor Segmentation

DRRNet: Dense Residual Refine Networks for Automatic Brain Tumor Segmentation

Adaptive Channel and Multiscale Spatial Context Network for Breast Mass Segmentation in Full-Field Mammograms

Brain tumor magnetic resonance image segmentation by a multiscale contextual attention module combined with a deep residual UNet (MCA-ResUNet)

A Dynamic Context Encoder Network for Liver Tumor Segmentation

Semantic segmentation of brain tumor with nested residual attention networks

Automatic cervical cancer segmentation in multimodal magnetic resonance imaging using an EfficientNet encoder in UNet++ architecture

Mixed 2D and 3D convolutional network with multi-scale context for lesion segmentation in breast DCE-MRI*

Three-dimensional affinity learning based multi-branch ensemble network for breast tumor segmentation in MRI

LCANet: A Lightweight Context-Aware Network for Bladder Tumor Segmentation in MRI Images

Three-dimensional breast tumor segmentation on DCE-MRI with a multilabel attention-guided joint-phase-learning network

DMAeEDNet: Dense Multiplicative Attention Enhanced Encoder Decoder Network for Ultrasound-Based Automated Breast Lesion Segmentation

Deep Learning Framework with Multi-Head Dilated Encoders for Enhanced Segmentation of Cervical Cancer on Multiparametric Magnetic Resonance Imaging

DECTNet: Dual Encoder Network combined convolution and Transformer architecture for medical image segmentation