无监督/半监督/对比学习

⚠️ 以下所有内容总结都来自于 大语言模型的能力，如有错误，仅供参考，谨慎使用
🔴 请注意：千万不要用于严肃的学术场景，只能用于论文阅读前的初筛！
💗 如果您觉得我们的项目对您有帮助 ChatPaperFree ，还请您给我们一些鼓励！⭐️ HuggingFace免费体验

2025-10-10 更新

The best performance in the CARE 2025 – Liver Task (LiSeg-Contrast): Contrast-Aware Semi-Supervised Segmentation with Domain Generalization and Test-Time Adaptation

Authors:Jincan Lou, Jingkun Chen, Haoquan Li, Hang Li, Wenjian Huang, Weihua Chen, Fan Wang, Jianguo Zhang

Accurate liver segmentation from contrast-enhanced MRI is essential for diagnosis, treatment planning, and disease monitoring. However, it remains challenging due to limited annotated data, heterogeneous enhancement protocols, and significant domain shifts across scanners and institutions. Traditional image-to-image translation frameworks have made great progress in domain generalization, but their application is not straightforward. For example, Pix2Pix requires image registration, and cycle-GAN cannot be integrated seamlessly into segmentation pipelines. Meanwhile, these methods are originally used to deal with cross-modality scenarios, and often introduce structural distortions and suffer from unstable training, which may pose drawbacks in our single-modality scenario. To address these challenges, we propose CoSSeg-TTA, a compact segmentation framework for the GED4 (Gd-EOB-DTPA enhanced hepatobiliary phase MRI) modality built upon nnU-Netv2 and enhanced with a semi-supervised mean teacher scheme to exploit large amounts of unlabeled volumes. A domain adaptation module, incorporating a randomized histogram-based style appearance transfer function and a trainable contrast-aware network, enriches domain diversity and mitigates cross-center variability. Furthermore, a continual test-time adaptation strategy is employed to improve robustness during inference. Extensive experiments demonstrate that our framework consistently outperforms the nnU-Netv2 baseline, achieving superior Dice score and Hausdorff Distance while exhibiting strong generalization to unseen domains under low-annotation conditions.

精确的肝脏分段对于对比增强MRI的诊断、治疗计划和疾病监测至关重要。然而，由于标注数据有限、增强协议存在异质性以及扫描仪和机构间的领域漂移等问题，它仍然是一个挑战。传统的图像到图像翻译框架在领域通用化方面取得了很大的进步，但应用并不直接。例如，Pix2Pix 需要图像注册，而 cycle-GAN 无法无缝集成到分段流水线中。同时，这些方法最初是用于处理跨模态场景的，经常会引入结构失真并面临训练不稳定的问题，这在我们单一的模态场景中可能会带来不利之处。为了解决这些挑战，我们提出了CoSSeg-TTA，这是一个紧凑的分段框架，适用于GED4（Gd-EOB-DTPA增强肝胆期MRI）模态，建立在nnU-Netv2之上，并用半监督教师方案进行增强以利用大量未标记的体积数据。领域适应模块结合了基于随机直方图的风格外观转换函数和可训练的对比感知网络，丰富了领域多样性并减轻了跨中心变异性。此外，采用连续测试时间适应策略来提高推理过程中的稳健性。大量实验表明，我们的框架始终优于nnU-Netv2基线，在Dice分数和Hausdorff距离上表现优越，同时在低注释条件下对未见领域具有很强的泛化能力。

论文及项目相关链接

PDF 11 pages, 3 figures

Summary

本文提出了一种基于nnU-Netv2的紧凑分割框架CoSSeg-TTA，用于对GED4（Gd-EOB-DTPA增强肝胆期MRI）模态进行精确肝脏分割。该框架结合了半监督均值教师方案，利用大量未标记体积数据，并通过域适应模块和持续测试时适应策略，解决了有限标注数据、异质增强协议和跨扫描仪和机构域漂移等问题，实现了出色的性能。

Key Takeaways

1. 准确肝脏分割在对比增强MRI中对诊断、治疗计划和疾病监测至关重要。
2. 传统图像到图像翻译框架在域推广方面取得了进展，但应用于单模态场景时存在挑战。
3. CoSSeg-TTA框架结合nnU-Netv2和半监督均值教师方案，能有效处理有限标注数据和异质增强协议问题。
4. 域适应模块通过随机直方图样式转换函数和可训练对比度感知网络，增强了域多样性和减轻了跨中心差异性。
5. 持续测试时适应策略提高了推理阶段的稳健性。
6. 实验表明，CoSSeg-TTA框架在未见域的条件下，相较于nnU-Netv2基线模型，实现了更高的Dice分数和Hausdorff距离，并表现出强大的泛化能力。

Cool Papers

点此查看论文截图

Contrastive-SDE: Guiding Stochastic Differential Equations with Contrastive Learning for Unpaired Image-to-Image Translation

Authors:Venkata Narendra Kotyada, Revanth Eranki, Nagesh Bhattu Sristy

Unpaired image-to-image translation involves learning mappings between source domain and target domain in the absence of aligned or corresponding samples. Score based diffusion models have demonstrated state-of-the-art performance in generative tasks. Their ability to approximate complex data distributions through stochastic differential equations (SDEs) enables them to generate high-fidelity and diverse outputs, making them particularly well-suited for unpaired I2I settings. In parallel, contrastive learning provides a powerful framework for learning semantic similarities without the need for explicit supervision or paired data. By pulling together representations of semantically similar samples and pushing apart dissimilar ones, contrastive methods are inherently aligned with the objectives of unpaired translation. Its ability to selectively enforce semantic consistency at the feature level makes contrastive learning particularly effective for guiding generation in unpaired scenarios. In this work, we propose a time-dependent contrastive learning approach where a model is trained with SimCLR by considering an image and its domain invarient feature as a positive pair, enabling the preservation of domain-invariant features and the discarding of domain-specific ones. The learned contrastive model then guides the inference of a pretrained SDE for the I2I translation task. We empirically compare Contrastive-SDE with several baselines across three common unpaired I2I tasks, using four metrics for evaluation. Constrastive-SDE achieves comparable results to the state-of-the-art on several metrics. Furthermore, we observe that our model converges significantly faster and requires no label supervision or classifier training, making it a more efficient alternative for this task.

无配对图像到图像的翻译学习是在没有对齐或相应样本的情况下，学习源域和目标域之间的映射。基于分数的扩散模型在生成任务中表现出了卓越的性能。它们通过随机微分方程（SDEs）来逼近复杂的数据分布，能够生成高保真和多样化的输出，使得它们特别适用于无配对I2I设置。同时，对比学习提供了一个强大的框架，用于学习语义相似性，而无需明确的监督或配对数据。通过拉近语义相似样本的表示，并推开不相似的样本，对比方法与无配对翻译的目标天然契合。其在特征层面选择性执行语义一致性的能力，使得对比学习在无配对场景中引导生成特别有效。在这项工作中，我们提出了一种时间依赖对比学习方法，该方法使用SimCLR训练模型，将图像及其域不变特征视为正样本对，从而保留域不变特征并丢弃域特定特征。学习得到的对比模型然后引导对预训练的SDE进行I2I翻译任务的推理。我们通过三个常见的无配对I2I任务和四个评估指标，对Contrastive-SDE与几个基准方法进行了实证比较。Contrastive-SDE在几个指标上实现了与最新技术相当的结果。此外，我们观察到我们的模型收敛得更快，且无需标签监督或分类器训练，使其成为此任务的一个更高效的替代方案。

论文及项目相关链接

PDF 9 pages, 3 figures

Summary

本文介绍了无配对图像到图像翻译任务中，基于分数扩散模型和对比学习的方法。文章指出，通过利用分数扩散模型的强大生成能力和对比学习的语义相似性学习能力，该方法在三项常见无配对图像翻译任务中实现了优异的表现。该模型不仅实现了与现有技术相当的结果，而且在收敛速度和无需标签监督或分类器训练方面表现出更高的效率。

Key Takeaways

1. 无配对图像到图像翻译任务介绍：重点阐述了在缺乏对应样本的情况下，源域和目标域之间的映射学习问题。
2. 分数扩散模型的应用：描述了分数扩散模型在生成任务中的先进性能，以及其通过随机微分方程（SDEs）近似复杂数据分布的能力。
3. 对比学习的应用：介绍了对比学习在无配对图像翻译任务中的重要性，及其在选择性执行语义一致性方面的作用。
4. 时间依赖性对比学习方法：提出了考虑图像和其域不变特征作为正样本对的时间依赖性对比学习方法，这有助于保留域不变特征并丢弃域特定特征。
5. 对比学习与SDE的结合：描述了将学习的对比模型用于引导预训练SDE进行图像翻译任务的方法。
6. 实证分析：文章在三个常见的无配对图像翻译任务中进行了实证比较，结果显示所提出的方法在多个评价指标上达到了与现有技术相当的结果。

Cool Papers

点此查看论文截图

Conditional Pseudo-Supervised Contrast for Data-Free Knowledge Distillation

Authors:Renrong Shao, Wei Zhang, Jun wang

Data-free knowledge distillation(DFKD) is an effective manner to solve model compression and transmission restrictions while retaining privacy protection, which has attracted extensive attention in recent years. Currently, the majority of existing methods utilize a generator to synthesize images to support the distillation. Although the current methods have achieved great success, there are still many issues to be explored. Firstly, the outstanding performance of supervised learning in deep learning drives us to explore a pseudo-supervised paradigm on DFKD. Secondly, current synthesized methods cannot distinguish the distributions of different categories of samples, thus producing ambiguous samples that may lead to an incorrect evaluation by the teacher. Besides, current methods cannot optimize the category-wise diversity samples, which will hinder the student model learning from diverse samples and further achieving better performance. In this paper, to address the above limitations, we propose a novel learning paradigm, i.e., conditional pseudo-supervised contrast for data-free knowledge distillation(CPSC-DFKD). The primary innovations of CPSC-DFKD are: (1) introducing a conditional generative adversarial network to synthesize category-specific diverse images for pseudo-supervised learning, (2) improving the modules of the generator to distinguish the distributions of different categories, and (3) proposing pseudo-supervised contrastive learning based on teacher and student views to enhance diversity. Comprehensive experiments on three commonly-used datasets validate the performance lift of both the student and generator brought by CPSC-DFKD. The code is available at https://github.com/RoryShao/CPSC-DFKD.git

无数据知识蒸馏（DFKD）是一种在保留隐私保护的同时解决模型压缩和传输限制的有效方法，近年来引起了广泛关注。目前，大多数现有方法都使用生成器来合成图像以支持蒸馏过程。尽管当前的方法已经取得了巨大的成功，但仍有许多问题有待探索。首先，深度学习中监督学习的出色性能促使我们探索DFKD上的伪监督范式。其次，当前的合成方法无法区分不同类别样本的分布，从而产生了可能导致教师评估错误的模糊样本。此外，当前的方法无法优化类别多样性的样本，这将阻碍学生模型从各种样本中学习，并进一步实现更好的性能。针对以上局限性，本文提出了一种新的学习范式，即基于条件伪监督对比的无数据知识蒸馏（CPSC-DFKD）。CPSC-DFKD的主要创新点包括：（1）引入条件生成对抗网络，为伪监督学习合成特定类别的多样化图像；（2）改进生成器的模块，以区分不同类别的分布；（3）提出基于教师和学生观点的伪监督对比学习以增强多样性。在三个常用数据集上的综合实验验证了CPSC-DFKD带来的学生和生成器的性能提升。代码可访问https://github.com/RoryShao/CPSC-DFKD.git。

论文及项目相关链接

PDF 13 pages

摘要
无监督数据下的知识蒸馏（DFKD）是解决模型压缩和传输限制同时保护隐私的有效方法，近年来备受关注。现有方法主要通过生成器合成图像进行知识蒸馏。本文提出一种新型学习范式——条件伪监督对比无监督数据知识蒸馏（CPSC-DFKD），以解决现有方法的问题。CPSC-DFKD的创新点包括：引入条件生成对抗网络进行伪监督学习，改进生成器模块以区分不同类别的分布，以及提出基于师生视角的伪监督对比学习以增强多样性。在三个常用数据集上的综合实验验证了CPSC-DFKD对学生模型和生成器的性能提升。

关键见解

1. 引入条件生成对抗网络进行伪监督学习，合成特定类别的多样图像。
2. 改进生成器模块，使其能够区分不同类别的样本分布。
3. 提出基于师生视角的伪监督对比学习，增强样本多样性。
4. CPSC-DFKD能提升学生和生成器的性能。
5. 该方法解决了现有DFKD方法在合成图像时的模糊性和类别区分问题。
6. CPSC-DFKD在三个常用数据集上进行了综合实验验证，表现出良好的性能。
7. 公开的代码资源便于其他研究者使用和学习。

Cool Papers

点此查看论文截图

Generalizing Supervised Contrastive learning: A Projection Perspective

Authors:Minoh Jeong, Alfred Hero

Self-supervised contrastive learning (SSCL) has emerged as a powerful paradigm for representation learning and has been studied from multiple perspectives, including mutual information and geometric viewpoints. However, supervised contrastive (SupCon) approaches have received comparatively little attention in this context: for instance, while InfoNCE used in SSCL is known to form a lower bound on mutual information (MI), the relationship between SupCon and MI remains unexplored. To address this gap, we introduce ProjNCE, a generalization of the InfoNCE loss that unifies supervised and self-supervised contrastive objectives by incorporating projection functions and an adjustment term for negative pairs. We prove that ProjNCE constitutes a valid MI bound and affords greater flexibility in selecting projection strategies for class embeddings. Building on this flexibility, we further explore the centroid-based class embeddings in SupCon by exploring a variety of projection methods. Extensive experiments on image and audio datasets demonstrate that ProjNCE consistently outperforms both SupCon and standard cross-entropy training. Our work thus refines SupCon along two complementary perspectives–information-theoretic and projection viewpoints–and offers broadly applicable improvements whenever SupCon serves as the foundational contrastive objective.

自监督对比学习（SSCL）作为一种强大的表征学习范式，已从多个角度进行了研究，包括互信息和几何观点。然而，在这种情况下，有监督对比（SupCon）方法受到了相对较少的关注：例如，虽然已知用于SSCL的InfoNCE形成互信息的下界，但SupCon与互信息之间的关系仍未被探索。为了弥补这一空白，我们引入了ProjNCE，它是InfoNCE损失的一种推广，通过引入投影函数和负对的调整项，统一了监督学习和自监督学习的对比目标。我们证明了ProjNCE构成了一个有效的互信息界，并为选择类嵌入的投影策略提供了更大的灵活性。基于这种灵活性，我们进一步探索了基于质心的类嵌入在SupCon中的应用，并尝试了各种投影方法。在图像和音频数据集上的大量实验表明，ProjNCE始终优于SupCon和标准交叉熵训练。因此，我们的工作从信息论和投影这两个互补的角度对SupCon进行了精炼，每当SupCon作为基本的对比目标时，都能提供广泛适用的改进。

论文及项目相关链接

PDF

Summary

本文研究了自监督对比学习（SSCL）中的监督对比（SupCon）方法，并指出其与互信息（MI）之间的关系尚未被探索。为此，文章引入了ProjNCE损失，它统一了监督和无监督对比目标，通过融入投影函数和对负样本的调整项，不仅构成有效的MI边界，而且为选择类嵌入的投影策略提供了更大的灵活性。实验表明，ProjNCE在图像和音频数据集上的表现均优于SupCon和标准交叉熵训练。

Key Takeaways

1. 监督对比（SupCon）在自监督对比学习（SSCL）中的研究尚未得到充分关注。
2. ProjNCE损失是对InfoNCE损失的推广，能够统一监督和无监督对比目标。
3. ProjNCE构成有效的互信息（MI）边界。
4. ProjNCE提供了在选择类嵌入的投影策略上的更大灵活性。
5. 通过实验验证了ProjNCE在图像和音频数据集上的表现优于SupCon和交叉熵训练。
6. 文章从信息论和投影两个角度完善了SupCon。

Cool Papers

点此查看论文截图