Click the card below to follow the " CVer " official account
AI/CV heavy dry goods, delivered in the first time
Click to enter —>【Medical Image Segmentation】WeChat Exchange Group
After being urged for a long time, CVer officially launched the CVPR 2023 paper inventory series ! Amusi has collected a total of 13 papers on medical image segmentation , which should be the latest and most comprehensive inventory of CVPR 2023 medical image segmentation on various platforms . Among them, semi-supervision occupies 5 articles! Up to now, the code links of 10 papers have been released (does not mean that they have been open sourced)!
If you want to know the latest and high-quality AI papers, practical projects, data sets, introductory overviews and a lot of learning materials, welcome to join CVer Computer Vision Knowledge Planet! It is updated continuously every day, I hope it will be helpful to you! Scan the QR code below to join the learning!
For more CVPR 2023 papers and open source code, see the link below:
https://github.com/amusi/CVPR2023-Papers-with-Code
CVPR 2023 Medical Image Segmentation Papers (13)
1. Label-free Liver Tumor Segmentation
Label-Free Liver Tumor Segmentation
Unit: Huazhong University of Science and Technology, The Chinese University of Hong Kong (Shenzhen), JHU, The First Affiliated Hospital of Nanjing Medical University
Paper: https://arxiv.org/abs/2303.14869
Code: https://github.com/MrGiovanni/SyntheticTumors
One sentence summary: This paper proposes an effective strategy for synthesizing liver tumors, and the code is open source!
2. DconnNet: Medical Image Segmentation Based on Oriented Connectivity
Directional Connectivity-based Segmentation of Medical Images
Unit: Duke University
Paper: https://arxiv.org/abs/2304.00145
Code: https://github.com/Zyun-Y/DconnNet
One sentence summary: DconnNet: a new type of directional connectivity modeling network for medical image segmentation, the core idea is to separate the directional subspace from the shared latent space, and use the extracted directional features to enhance the overall data representation, performance outstanding! Better than nnU-Net and other networks, the code has been open source!
3. BCP: Bidirectional copy-paste for semi-supervised medical image segmentation
Bidirectional Copy-Paste for Semi-Supervised Medical Image Segmentation
Unit: East China Normal University, Shanghai Jiao Tong University
Paper: https://arxiv.org/abs/2305.00673
Code: https://github.com/DeepMed-Lab-ECNU/BCP
One sentence summary: This paper proposes a bidirectional copy-paste (BCP) method for semi-supervised medical image segmentation, which is simple and efficient! It can significantly improve the performance of the existing segmentation model, such as helping the SS-Net network increase, the code has been open source!
4. Mask Transformers for real-world medical image segmentation and out-of-distribution localization
Devil is in the Queries: Advancing Mask Transformers for Real-world Medical Image Segmentation and Out-of-Distribution Localization
Unit: Ali, Peking University, Guangdong Provincial People's Hospital, Shengjing Hospital, etc.
Paper: https://arxiv.org/abs/2304.00212
Code: None
One-sentence summary: It is said that this is the first work to explore the near-OOD detection and localization problem in medical image segmentation, in which MaxQuery network and QD loss are proposed, and the performance is SOTA!
5. FedCE: Fair Federated Medical Image Segmentation Based on Client Contribution Estimation
Fair Federated Medical Image Segmentation via Client Contribution Estimation
Unit: CUHK, NVIDIA
Paper: https://arxiv.org/abs/2303.16520
Code: https://github.com/NVIDIA/NVFlare/tree/dev/research/fed-ce
One-sentence summary: FedCE: A new federated learning method for medical image segmentation that uses client-side contribution estimates as global model aggregation weights, empirically evaluated on two real-world medical datasets, with significant performance improvements, more Good collaborative fairness, better performance fairness.
6. Ambiguous medical image segmentation based on diffusion model
Ambiguous Medical Image Segmentation using Diffusion Models
Unit: JHU, University of British Columbia
Homepage: https://aimansnigdha.github.io/cimd/
Paper: https://arxiv.org/abs/2304.04745
Code: https://github.com/aimansnigdha/Ambiguous-Medical-Image-Segmentation-using-Diffusion-Models
One-sentence summary: Validation is verified on three different medical image modalities (CT, ultrasound, and MRI), and a novel metric is also proposed to evaluate the diversity and accuracy of segmentation predictions, which is consistent with the clinical Practice, the code is open source!
7. Orthogonal labeling is beneficial to Barely supervised medical image segmentation
Orthogonal Annotation Benefits Barely-supervised Medical Image Segmentation
Units: Nanjing University (Shi Yinghuan's team), Southeast University, Shandong Women's University
Paper: https://arxiv.org/abs/2303.13090
Code: https://github.com/HengCai-NJU/DeSCO
One-sentence summary: This paper proposes a new annotation method for 3D medical image segmentation: orthogonal annotation, that is, marking two orthogonal slices for a volume, which greatly reduces the burden of annotation, and proposes DeSCO: Dense-Sparse Joint Training Paradigm , the segmentation performance is excellent!
8. MagicNet: Semi-Supervised Multi-Organ Segmentation for Magic-Cube Partitioning and Restoration
MagicNet: Semi-Supervised Multi-Organ Segmentation via Magic-Cube Partition and Recovery
Units: East China Normal University, Shanghai Jiaotong University, HKU
Paper: https://arxiv.org/abs/2301.01767
Code: https://github.com/DeepMed-Lab-ECNU/MagicNet
One-sentence summary: Demonstrated the effectiveness of MagicNet on two public CT multi-organ datasets, significantly outperforming state-of-the-art semi-supervised medical image segmentation methods, with a DSC improvement of +7 on the MACT dataset with 10% labeled images %
9. MCF: A Mutual Correction Framework for Semi-Supervised Medical Image Segmentation
MCF: Mutual Correction Framework for Semi-Supervised Medical Image Segmentation
Unit: Chongqing University of Posts and Telecommunications
Paper: https://openaccess.thecvf.com/content/CVPR2023/html/Wang_MCF_Mutual_Correction_Framework_for_Semi-Supervised_Medical_Image_Segmentation_CVPR_2023_paper.html
Code: https://github.com/WYC-321/MCF
One sentence summary: This paper discusses the problem of model bias correction, and proposes a new framework for semi-supervised medical image segmentation: MCF, with excellent performance!
10. Rethinking small-sample medical segmentation: a vector quantization perspective
Rethinking Few-Shot Medical Segmentation: A Vector Quantization View
Unit: Beijing Institute of Technology
Paper: https://openaccess.thecvf.com/content/CVPR2023/html/Huang_Rethinking_Few-Shot_Medical_Segmentation_A_Vector_Quantization_View_CVPR_2023_paper.html
Code: None
One-sentence summary: The VQ framework yields state-of-the-art performance on abdominal, cardiac, and prostate MRI datasets, and it is expected that this work will trigger a rethinking of the design of current small-shot medical segmentation models.
11. Pseudo-label-guided contrastive learning for semi-supervised medical image segmentation
Pseudo-label Guided Contrastive Learning for Semi-supervised Medical Image Segmentation
Unit: Stony Brook University
Paper: https://openaccess.thecvf.com/content/CVPR2023/html/Basak_Pseudo-Label_Guided_Contrastive_Learning_for_Semi-Supervised_Medical_Image_Segmentation_CVPR_2023_paper.html
Code: https://github.com/hritam-98/PatchCL-MedSeg
One sentence summary: According to the author, this is the first attempt to integrate contrastive learning in a semi-supervised environment using consistency regularization and pseudo-labels for semi-supervised medical image segmentation. The performance is excellent, and the code is open source!
12. SDC-UDA: A Volumetric Unsupervised Domain Adaptation Framework for Cross-Modal Medical Image Segmentation
SDC-UDA: Volumetric Unsupervised Domain Adaptation Framework for Slice-Direction Continuous Cross-Modality Medical Image Segmentation
Unit: Yonsei University, Naver AI Lab, Harvard Medical School, etc.
Paper: https://arxiv.org/abs/2305.11012
Code: None
One sentence summary: SDC-UDA: A new volumetric unsupervised domain adaptation framework for slice-oriented continuous cross-modal medical image segmentation, and its effectiveness has been verified on multiple public datasets, achieving the most advanced Split performance.
13. DoNet: Deep De-overlapping Networks for Cytological Instance Segmentation
DoNet: Deep De-overlapping Network for Cytology Instance Segmentation
Unit: HKUST, Tencent AI Lab
Paper: https://arxiv.org/abs/2303.14373
Code: https://github.com/DeepDoNet/DoNet
One sentence summary: DoNet: a de-overlapping network based on decomposition and reorganization strategies for cell instance segmentation, performance SOTA on ISBI2014 and CPS data sets!
If you want to know the latest and high-quality AI papers, practical projects, data sets, introductory overviews and a lot of learning materials, welcome to join CVer Computer Vision Knowledge Planet! It is updated continuously every day, I hope it will be helpful to you! Scan the QR code below to join the learning!
上面13篇医学图像分割论文和论文下载
CVer后台回复：CVPR2023，即可下载论文和代码Medical image and image segmentation exchange group established
扫描下方二维码，或者添加微信：CVer333，即可添加CVer小助手微信，便可申请加入CVer-医学图像或者图像分割 微信交流群。另外其他垂直方向已涵盖：目标检测、图像分割、目标跟踪、人脸检测&识别、OCR、姿态估计、超分辨率、SLAM、医疗影像、Re-ID、GAN、NAS、深度估计、自动驾驶、强化学习、车道线检测、模型剪枝&压缩、去噪、去雾、去雨、风格迁移、遥感图像、行为识别、视频理解、图像融合、图像检索、论文投稿&交流、PyTorch、TensorFlow和Transformer等。
一定要备注：研究方向+地点+学校/公司+昵称（如医学图像或者图像分割+上海+上交+卡卡），根据格式备注，可更快被通过且邀请进群
▲扫码或加微信号: CVer333，进交流群
CVer计算机视觉（知识星球）来了！想要了解最新最快最好的CV/DL/AI论文速递、优质实战项目、AI行业前沿、从入门到精通学习教程等资料，欢迎扫描下方二维码，加入CVer计算机视觉，已汇集数千人！
▲扫码进星球
▲点击上方卡片，关注CVer公众号It's not easy to organize, please like and watch