A Tumor Segmentation Method Based on Mean-Teacher Reusing Pseudo-Labels DOI Creative Commons
Chengyu Jiang,

Shangkun Liu,

Jingyu Wang

et al.

IEEE Access, Journal Year: 2024, Volume and Issue: 12, P. 41942 - 41953

Published: Jan. 1, 2024

Breast tumor is a common female physiological disease, and the malignant one of main fatal diseases women. Accurate examination assessment shape can facilitate subsequent treatment improve cure rate. With development deep learning, automatic detection systems are designed to assist doctors in diagnosis. However, blurry edges, poor visual quality, irregular shapes breast tumors pose significant challenges design highly efficient system. In addition, lack publicly available labeled data major obstacle developing accurate robust learning models for detection. To overcome aforementioned issues, we propose SRU-PMT+, pseudo-label reusing Mean-Teacher architecture based on squeeze-and-excitation residual (SE-Res) attention. We utilize proposed segmentation network, SRU-Net++, generate pseudo-labels unlabeled data, guide student model using generated groundtruth, improving accuracy robustness model. Our semi-supervised method has been rigorously evaluated dataset, i.e., Ultrasound Images (BUSI) dataset. Results show that our outperforms current methods good performance. Importantly, strategy improves performance segmentation.

Language: Английский

Explainable attention based breast tumor segmentation using a combination of UNet, ResNet, DenseNet, and EfficientNet models DOI Creative Commons
Shokofeh Anari, Soroush Sadeghi, Ghazaal Sheikhi

et al.

Scientific Reports, Journal Year: 2025, Volume and Issue: 15(1)

Published: Jan. 6, 2025

This study utilizes the Breast Ultrasound Image (BUSI) dataset to present a deep learning technique for breast tumor segmentation based on modified UNet architecture. To improve accuracy, model integrates attention mechanisms, such as Convolutional Block Attention Module (CBAM) and Non-Local Attention, with advanced encoder architectures, including ResNet, DenseNet, EfficientNet. These mechanisms enable focus more effectively relevant areas, resulting in significant performance improvements. Models incorporating outperformed those without, reflected superior evaluation metrics. The effects of Dice Loss Binary Cross-Entropy (BCE) model's were also analyzed. maximized overlap between predicted actual masks, leading precise boundary delineation, while BCE achieved higher recall, improving detection areas. Grad-CAM visualizations further demonstrated that attention-based models enhanced interpretability by accurately highlighting findings denote combining framework can yield reliable accurate segmentation. Future research will explore use multi-modal imaging, real-time deployment clinical applications, performance.

Language: Английский

Citations

6
Competitive dual-students using bi-level contrastive learning for semi-supervised medical image segmentation DOI
Gang Hu, Zhao Feng, Essam H. Houssein

et al.

Engineering Applications of Artificial Intelligence, Journal Year: 2025, Volume and Issue: 144, P. 110082 - 110082

Published: Jan. 26, 2025

Language: Английский

Citations

1
MSDANet: A multi-scale dilation attention network for medical image segmentation DOI
Jinquan Zhang,

Zhuang Luan,

Lina Ni

et al.

Biomedical Signal Processing and Control, Journal Year: 2023, Volume and Issue: 90, P. 105889 - 105889

Published: Dec. 28, 2023

Language: Английский

Citations

21
Swin-Net: A Swin-Transformer-Based Network Combing with Multi-Scale Features for Segmentation of Breast Tumor Ultrasound Images DOI Creative Commons
Chengzhang Zhu,

Xian Chai,

Yalong Xiao

et al.

Diagnostics, Journal Year: 2024, Volume and Issue: 14(3), P. 269 - 269

Published: Jan. 26, 2024

Breast cancer is one of the most common cancers in world, especially among women. tumor segmentation a key step identification and localization breast region, which has important clinical significance. Inspired by swin-transformer model with powerful global modeling ability, we propose semantic framework named Swin-Net for ultrasound images, combines Transformer Convolutional Neural Networks (CNNs) to effectively improve accuracy segmentation. Firstly, our utilizes encoder stronger learning can extract features images more precisely. In addition, two new modules are introduced method, including feature refinement enhancement module (RLM) hierarchical multi-scale fusion (HFM), given that influence ultrasonic image acquisition methods characteristics lesions difficult capture. Among them, RLM used further refine enhance map learned transformer encoder. The HFM process high-level low-level details, so as achieve effective cross-layer fusion, suppress noise, performance. Experimental results show performs significantly better than advanced on public benchmark datasets. particular, it achieves an absolute improvement 1.4–1.8% Dice. Additionally, provide dataset test effect model, demonstrating validity method. summary, proposed makes significant advancements segmentation, providing valuable exploration research applications this domain.

Language: Английский

Citations

8
Segmenting medical images with limited data DOI
Zhaoshan Liu, Qiujie Lv, Chau Hung Lee

et al.

Neural Networks, Journal Year: 2024, Volume and Issue: 177, P. 106367 - 106367

Published: May 7, 2024

Language: Английский

Citations

6
NFMPAtt-Unet: Neighborhood Fuzzy C-means Multi-scale Pyramid Hybrid Attention Unet for medical image segmentation DOI
Xinpeng Zhao,

Weihua Xu

Neural Networks, Journal Year: 2024, Volume and Issue: 178, P. 106489 - 106489

Published: June 22, 2024

Language: Английский

Citations

5
PH-Net: Semi-Supervised Breast Lesion Segmentation via Patch-Wise Hardness DOI

Siyao Jiang,

Huisi Wu, Junyang Chen

et al.

2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Journal Year: 2024, Volume and Issue: 30, P. 11418 - 11427

Published: June 16, 2024

Language: Английский

Citations

5
EfficientUNetViT: Efficient Breast Tumor Segmentation Utilizing UNet Architecture and Pretrained Vision Transformer DOI Creative Commons
Shokofeh Anari, Gabriel Gomes de Oliveira, Ramin Ranjbarzadeh

et al.

Bioengineering, Journal Year: 2024, Volume and Issue: 11(9), P. 945 - 945

Published: Sept. 21, 2024

This study introduces a sophisticated neural network structure for segmenting breast tumors. It achieves this by combining pretrained Vision Transformer (ViT) model with UNet framework. The architecture, commonly employed biomedical image segmentation, is further enhanced depthwise separable convolutional blocks to decrease computational complexity and parameter count, resulting in better efficiency less overfitting. ViT, renowned its robust feature extraction capabilities utilizing self-attention processes, efficiently captures the overall context within images, surpassing performance of conventional networks. By using ViT as encoder our model, we take advantage extensive representations acquired from datasets, major enhancement model’s ability generalize train efficiently. suggested has exceptional cancers medical highlighting advantages integrating transformer-based encoders efficient topologies. hybrid methodology emphasizes transformers field processing establishes new standard accuracy activities related tumor segmentation.

Language: Английский

Citations

5
LRseg: An efficient railway region extraction method based on lightweight encoder and self-correcting decoder DOI
Zhicheng Feng, Jie Yang, Zhichao Chen

et al.

Expert Systems with Applications, Journal Year: 2023, Volume and Issue: 238, P. 122386 - 122386

Published: Nov. 1, 2023

Language: Английский

Citations

11
Memory-efficient transformer network with feature fusion for breast tumor segmentation and classification task DOI
Ahmed Iqbal, Muhammad Sharif

Engineering Applications of Artificial Intelligence, Journal Year: 2023, Volume and Issue: 127, P. 107292 - 107292

Published: Nov. 11, 2023

Language: Английский

Citations

11