Scene labeling is a challenging classification problem where each input image requires a pixel-level prediction map. Recently, deep-learning-based methods have shown their effectiveness on solving this problem. However, we argue that the large intra-class variation provides ambiguous training information and hinders the deep models' ability to learn more discriminative deep feature representations. Unlike existing methods that mainly utilize semantic context for regularizing or smoothing the prediction map, we design novel supervisions from semantic context for learning better deep feature representations. Two types of semantic context, scene names of images and label map statistics of image patches, are exploited to create label hierarchies between the original classes and newly created subclasses as the learning supervisions. Such subclasses show lower intra-class variation, and help CNN detect more meaningful visual patterns and learn more effective deep features. Novel training strategies and network structure that take advantages of such label hierarchies are introduced. Our proposed method is evaluated extensively on four popular datasets, Stanford Background (8 classes), SIFTFlow (33 classes), Barcelona (170 classes) and LM+Sun datasets (232 classes) with 3 different networks structures, and show state-of-the-art performance. The experiments show that our proposed method makes deep models learn more discriminative feature representations without increasing model size or complexity.
Real-World Point Tracking with Verifier-Guided Pseudo-Labeling
Department of Computer Engineering, Koç University（科卡大学计算机工程系）；KUIS AI Center（KUIS人工智能中心）；School of Artificial Intelligence, Shanghai Jiao Tong University（上海交通大学人工智能学院）
Decoupled Pseudo-labeling for Semi-Supervised Monocular 3D Object Detection
School of Computer Science and Engineering, Sun Yat-sen University（中山大学计算机科学与工程学院）；Baidu Inc.（百度公司）；GuangDong Province Key Laboratory of Information Security Technology（广东省信息安全技术重点实验室）
Iterative pseudo-labeling based adaptive copy-paste supervision for semi-supervised tumor segmentation
School of Software, Northwestern Polytechnical University, Shaanxi, China（西北工业大学软件学院）；Yangtze River Delta Research Institute of Northwestern Polytechnical University, Taicang, China（西北工业大学 Yangtze River Delta 研究院）；Department of Computer Science and Information Technology, La Trobe University, Melbourne, Australia（拉特罗布大学计算机科学与信息技术系）；Department of Computer Science, School of Engineering, Shantou University, Guangdong, China（汕尾大学工程学院计算机科学系）；Shandong Cancer Hospital and Institute, Shandong First Medical University and Shandong Academy of Medical Sciences, Shandong, China（山东省肿瘤医院）；Australian Institute of Health Innovation (AIHI), Macquarie University, Australia（澳大利亚健康创新研究所（AIHI），麦考瑞大学）；School of Software, Shandong University, Shandong, China（山东大学软件学院）；AIDD, Faculty of Applied Science, Macao Polytechnic University, Macao SAR, China（澳门理工学院应用科学学院AIDD）；School of Computer Software, College of Intelligence and Computing, Tianjin University, Tianjin, China（天津大学计算机软件学院）
为了从人群图像中学习可靠的人计数器，通常需要头部中心注释。然而，在密集的人群中注释头部中心是一个费力和乏味的过程。在本文中，我们提出了一个主动学习框架，它可以在有限的监督下实现准确的人群计数：给定一个小的标记预算，而不是随机选择图像进行注释，我们首先引入一个主动标记策略来注释数据集中信息量最大的图像，并在此基础上学习计数模型。重复该过程，使得在每个循环中，我们选择人群密度不同且与先前选择不同的样本。在最后一个循环中，当标签预算得到满足时，大量的未标记数据也被利用：引入分布分类器来对齐标记数据和未标记数据;此外，我们建议混合网络中数据的分布标签和潜在表示，以特别改善训练样本之间的分布对齐。我们遵循流行的密度估计管道进行人群计数。在上海理工大学、UCF CC 50、MAll、TRANCOS和DCC等标准基准上进行了大量的实验。通过注释有限数量的图像（例如，数据集的10%），我们的方法达到的性能水平与利用数据集的完整注释的最新技术水平相差不远。
Foundation Model-guided Iteratively Prompting and Pseudo-Labeling for Partially Labeled Medical Image Segmentation
Department of Biomedical Engineering, Columbia University（哥伦比亚大学生物医学工程系）；Department of Radiation Oncology, Columbia University Irving Medical Center（哥伦比亚大学欧文医学中心放射肿瘤科）；Data Science Institute, Columbia University（哥伦比亚大学数据科学研究所）；Herbert Irving Comprehensive Cancer Center, Columbia University（哥伦比亚大学赫伯特·欧文综合癌症中心）；Department of Radiation Oncology, Icahn School of Medicine at Mount Sinai（西奈山伊坎医学院放射肿瘤科）
Principle-Guided Supervision for Interpretable Uncertainty in Medical Image Segmentation
School of Data Science, Fudan University（复旦大学数据科学学院）；Institute of Science and Technology for Brain-Inspired Intelligence, Fudan University（复旦大学脑启发智能科学研究院）；National Heart and Lung Institute, Imperial College London（伦敦帝国理工学院国家心脏和肺研究所）
人的重新识别（re-ID）旨在检索不同相机拍摄的同一个人的图像。re-ID的一个挑战是当模型用于与训练数据域（源数据）属于不同域的感兴趣数据（目标数据）时的性能保持。无监督领域自适应（UDA）是一个有趣的研究方向，因为它避免了昂贵的目标数据的注释。伪标记方法在基于UDA的re-ID中获得了最好的结果。令人惊讶的是，标记的源数据在该初始化步骤之后被丢弃。然而，我们认为伪标记可以进一步利用标记的源数据，以改善初始化后的训练步骤。为了提高对错误伪标签的鲁棒性，我们提倡在所有训练迭代期间利用标记的源数据和伪标记的目标数据。为了支持我们的指导方针，我们引入了一个框架，它依赖于一个两个分支架构优化分类和三元组损失为基础的度量学习，分别在源和目标域，以允许\emph{adaptability to the target domain}，同时确保\emph{robustness to noisy pseudo-labels}。事实上，共享的低和中级别参数受益于源分类和三重丢失信号，而目标分支的高级别参数学习特定于域的特征。我们的方法很简单，可以很容易地与现有的伪标记UDA方法相结合。我们的实验表明，它是有效的，并提高了性能时，基本方法没有机制来处理伪标签噪声或硬适应任务。我们的方法在常用数据集Market-1501和DukeMTMC-reID上进行评估时达到了最先进的性能，并且在针对更大，更具挑战…
Dense Supervision Propagation for Weakly Supervised Semantic Segmentation on 3D Point Clouds
School of Electrical and Electronic Engineering, Nanyang Technological University（南洋理工大学电气与电子工程学院）；School of Computer Science and Engineering, Nanyang Technological University（南洋理工大学计算机科学与工程学院）；Delta Research Center（三角洲研究中心）
Pseudo-label Guided Cross-video Pixel Contrast for Robotic Surgical Scene Segmentation with Limited Annotations
The Chinese University of Hong Kong（香港中文大学）；Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences（中国科学院深圳先进技术研究院）；University College London（伦敦大学学院）；Zhejiang Lab（之江实验室）
Dual-Thresholded Heatmap-Guided Proposal Clustering and Negative Certainty Supervision with Enhanced Base Network for Weakly Supervised Object Detection
Institute of Cyberspace Security, Harbin Institute of Technology（哈尔滨工业大学网络安全学院）；Faculty of Information Technology, Monash University（莫纳什大学信息科技学院）；Center on Machine Learning Research, Harbin Institute of Technology（哈尔滨工业大学机器学习研究中心）；Department of New Networks, Peng Cheng Laboratory（鹏城实验室网络部）；School of Cyberspace Science, Harbin Institute of Technology（哈尔滨工业大学网络空间科学学院）
Weakly-Supervised Concealed Object Segmentation with SAM-based Pseudo Labeling and Multi-scale Feature Grouping
Tsinghua University（清华大学）；NEC Laboratories America（NEC美国实验室）；Nanjing University of Posts and Telecommunications（南京邮电大学）；ETH Zürich（苏黎世联邦理工学院）；Tianyi Traffic Technology（天翼交通科技）
针对弱监督伪装目标分割中内在相似性和弱监督的挑战，提出基于多尺度特征分组模块和 SAM 伪标签生成及优化策略的新方法，实现了最先进的分割性能。
Wangxuan Institute of Computer Technology, Peking University（北京大学计算机技术研究院）；Chongqing Changan Automobile Co., Ltd（重庆长安汽车有限公司）；University of California, Merced（加州大学默塞德分校）
