RANet: Ranking Attention Network for Fast Video Object Segmentation
The University of Sydney(悉尼大学);Institute of Artificial Intelligence and Robotics, Xi’an Jiaotong University(西安交通大学人工智能与机器人研究所);Inception Institute of Artificial Intelligence (IIAI)(起源人工智能研究院);College of Computer Science, Nankai University(南开大学计算机学院)
Despite online learning (OL) techniques have boosted the performance of semi-supervised video object segmentation (VOS) methods, the huge time costs of OL greatly restrict their practicality. Matching based and propagation based methods run at a faster speed by avoiding OL techniques. However, they are limited by sub-optimal accuracy, due to mismatching and drifting problems. In this paper, we develop a real-time yet very accurate Ranking Attention Network (RANet) for VOS. Specifically, to integrate the insights of matching based and propagation based methods, we employ an encoder-decoder framework to learn pixel-level similarity and segmentation in an end-to-end manner. To better utilize the similarity maps, we propose a novel ranking attention module, which automatically ranks and selects these maps for fine-grained VOS performance. Experiments on DAVIS-16 and DAVIS-17 datasets show that our RANet achieves the best speed-accuracy trade-off, e.g., with 33 milliseconds per frame and J&F=85.5% on DAVIS-16. With OL, our RANet reaches J&F=87.1% on DAVIS-16, exceeding state-of-the-art VOS methods. The code can be found at https://github.com/Storife/RANet.
On The Usage Of Average Hausdorff Distance For Segmentation Performance Assessment: Hidden Bias When Used For Ranking
CLAIM - Charité Lab for AI in Medicine, Charité Universitätsmedizin Berlin(柏林夏里特医学院人工智能医学实验室);Research Studio Data Science, Research Studios Austria(奥地利研究工作室数据科学研究工作室);Centre for Stroke Research Berlin, Charité Universitätsmedizin Berlin(柏林夏里特医学院柏林卒中研究中心);Department of Neurology, Max Planck Institute for Human Cognitive and Brain Sciences(马克斯·普朗克人类认知与脑科学研究所神经病学系);School of Computing and Digital Technology, Faculty of Computing, Engineering and the Built Environment, Birmingham City University(伯明翰城市大学计算、工程与建筑环境学院计算与数字技术系);Mind, Brain, Body Institute, Berlin School of Mind and Brain, Humboldt-Universität Berlin(柏林洪堡大学柏林心智与脑学院心智、脑与身体研究所)
Data-Centric Benchmark for Label Noise Estimation and Ranking in Remote Sensing Binary Building Segmentation
School of Computer Science and Informatics, University of Liverpool(利物浦大学计算机科学与信息学学院);German Aerospace Center (DLR), Remote Sensing Technology Institute(德国航天中心(DLR)遥感技术研究所);Division of Geoinformatics, KTH Royal Institute of Technology(皇家理工学院地理信息系);University College London(伦敦大学学院);University of Exeter(埃克塞特大学);Inria(法国国家信息与自动化技术研究院);German Aerospace Center (DLR), Microwaves and Radar Institute(德国航天中心(DLR)微波与雷达研究所)
NucFuseRank: Dataset Fusion and Performance Ranking for Nuclei Instance Segmentation
Research Center for Medical Image Analysis and Artificial Intelligence, Department of Medicine, Faculty of Medicine and Dentistry, Danube Private University(医学图像分析与人工智能研究中心,医学系,医学与牙科学院,丹ube私立大学);Department of Pathophysiology and Allergy Research, Center of Pathophysiology, Infectiology and Immunology, Medical University of Vienna(病理生理学与过敏研究部门,病理学、感染学与免疫学中心,维也纳医学大学);Comprehensive Center for AI in Medicine (CAIM), Medical University of Vienna(医学人工智能综合中心(CAIM),维也纳医学大学);Research Center for Clinical AI-Research in Omics and Medical Data Science, Department of Medicine, Faculty of Medicine and Dentistry, Danube Private University(临床AI研究在组学和医学数据科学研究中心,医学系,医学与牙科学院,丹ube私立大学);Austrian Center for Medical Innovation and Technology(奥地利医学创新与技术中心)
Automated segmentation of pediatric neuroblastoma on multi-modal MRI: Results of the SPPIN challenge at MICCAI 2023
Princess Máxima Center for Pediatric Oncology(玛西玛公主儿科肿瘤中心);University Medical Center Utrecht(乌得勒支大学医学中心);Utrecht University(乌得勒支大学);Osaka University(大阪大学);Muroran Institute of Technology(室兰工业大学);Niigata University of Health and Welfare(新潟医疗福祉大学);AGH University of Krakow(克拉科夫AGH科技大学);HES-SO Valais-Wallis(瑞士西部应用科学与艺术大学瓦莱分校);University of Geneva(日内瓦大学);Shanghai Artificial Intelligence Laboratory(上海人工智能实验室);South China Normal University(华南师范大学);La Fe Health Research Institute(拉费健康研究所);University of Nebraska Medical Center(内布拉斯加大学医学中心);Wageningen Food Safety Research(瓦赫宁根食品安全研究所)
Clinical Interpretability of Deep Learning Segmentation Through Shapley-Derived Agreement and Uncertainty Metrics
Department of Mechanical Engineering, University of Washington(华盛顿大学机械工程系);University of Washington School of Medicine(华盛顿大学医学院);Department of Neurological Surgery, University of Washington(华盛顿大学神经外科系)
Challenges in Deep Learning-Based Small Organ Segmentation: A Benchmarking Perspective for Medical Research with Limited Datasets
Centre for Industrial Software, University of Southern Denmark(南丹麦大学工业软件中心);Centre for Industrial Mechanics, University of Southern Denmark(南丹麦大学工业力学中心);Duke-NUS(杜克-新加坡国立大学医学院);National Heart Center Singapore(新加坡国家心脏中心);University of Medicine and Pharmacy Carol Davila Bucharest(布加勒斯特卡罗尔·达维拉医药大学)
Zhejiang University(浙江大学);Megvii Inc. (Face++)(旷视科技有限公司(Face++));Huazhong University of Science and Technology(华中科技大学);Peking University(北京大学);The University of Tokyo(东京大学)
Maximum Matching Accuracy: An Instance Segmentation Evaluation Metric Utilizing Globally Optimal Matching
College of Computing, Georgia Institute of Technology(佐治亚理工学院计算机学院);School of Materials Science and Engineering, Georgia Institute of Technology(佐治亚理工学院材料科学与工程学院);Wallace H. Coulter Department of Biomedical Engineering, Georgia Institute of Technology(佐治亚理工学院生物医学工程系);George W. Woodruff School of Mechanical Engineering, Georgia Institute of Technology(佐治亚理工学院机械工程学院)
2nd Place Solution for MeViS Track in CVPR 2024 PVUW Workshop: Motion Expression guided Video Segmentation
Institute of Automation, Chinese Academy of Sciences (CASIA)(中国科学院自动化研究所);School of Artificial Intelligence, University of Chinese Academy of Sciences (UCAS)(中国科学院大学人工智能学院);Beijing Academy of Artificial Intelligence (BAAI)(北京人工智能研究院);University of Science and Technology Beijing (USTB)(北京科技大学)
SegWithU: Uncertainty as Perturbation Energy for Single-Forward-Pass Risk-Aware Medical Image Segmentation
University of Toronto(多伦多大学);McGill University(麦吉尔大学);University of Waterloo(滑铁卢大学);Vector Institute(向量研究所);Project Neura;University of Toronto Machine Intelligence Student Team(多伦多大学人工智能学生团队);Amplimit
Deep CNN ensembles and suggestive annotations for infant brain MRI segmentation
Ecole de Technologie Superieure (ETS), University of Quebec(魁北克大学高等技术学院);University of North Carolina at Chapel Hill(北卡罗来纳大学教堂山分校);Xidian University(西安电子科技大学);Korea University(高丽大学)
GeoSpark: Sparking up Point Cloud Segmentation with Geometry Clue
University of Cambridge(剑桥大学);HKUST(GuangZhou)(香港科技大学(广州))
当前的点云分割架构受到有限的长距离特征建模的影响,因为它们主要依赖于与局部邻域的信息聚合。此外,为了在多个尺度上学习点特征,大多数方法利用数据不可知的采样方法来减少每个阶段之后的点的数量。然而,这种采样方法通常在早期阶段丢弃小对象的点,导致特征学习不足。我们相信这些问题可以通过引入明确的几何线索作为指导来缓解。为此,我们提出了GeoSpark,这是一个插件模块,它将几何线索整合到网络中,以激发特征学习和下采样。GeoSpark可以轻松集成到各种主干中。对于特征聚合,它通过允许网络从局部点和相邻几何分区中学习来改进特征建模,从而扩大数据定制的感受野。此外,GeoSpark利用几何分区信息来指导下采样过程,其中保留具有独特特征的点,同时融合冗余点,从而更好地保留整个网络中的关键点。在将GeoSpark添加到包括PointNet++、KPConv和PointTransformer在内的各种主干之后,我们观察到了一致的改进。值得注意的是,当与Point Transformer集成时,我们的GeoSpark模块在ScanNetv2数据集上实现了74.7%的mIoU(提高了4.1%),在S3DIS Area 5数据集上实现了71.5%的mIoU(提高了1.1%),在两个基准测试中都名列前茅。代码和模型将公开发布。
Semi-supervised Cardiac Image Segmentation via Label Propagation and Style Transfer
Institute of Computing Technology, Chinese Academy of Sciences(中国科学院计算技术研究所);University of Chinese Academy of Sciences(中国科学院大学);University of California, Los Angeles(加州大学洛杉矶分校);Lenovo Research(联想研究院);Lenovo Ltd.(联想有限公司)
QU-BraTS: MICCAI BraTS 2020 Challenge on Quantifying Uncertainty in Brain Tumor Segmentation - Analysis of Ranking Scores and Benchmarking Results
McGill University(麦吉尔大学);University of Oxford(牛津大学);University of Pennsylvania(宾夕法尼亚大学);Perelman School of Medicine at the University of Pennsylvania(宾夕法尼亚大学佩雷尔曼医学院);University of Bern(伯尔尼大学);Bern University Hospital(伯尔尼大学医院);University of Texas Southwestern Medical Center(德克萨斯大学西南医学中心);Universidad de los Andes(洛斯安第斯大学);Imperial College London(帝国理工学院);Indian Statistical Institute(印度统计研究所);University of Calcutta(加尔各答大学);Uppsala University(乌普萨拉大学);The University of Pittsburgh Medical Center(匹兹堡大学医学中心);Inserm(法国国家健康与医学研究院);University of Auckland(奥克兰大学);Auckland City Hospital(奥克兰市医院);Massachusetts General Hospital(麻省总医院);Massachusetts Institute of Technology(麻省理工学院);EPITA Research and Development Laboratory (LRDE)(EPITA研发实验室);Old Dominion University(老道明大学);University of Girona(赫罗纳大学);University of Tehran(德黑兰大学);University of Nottingham(诺丁汉大学);NVIDIA(英伟达);University of Virginia(弗吉尼亚大学);Indian Institute of Technology - Jodhpur(印度理工学院焦特布尔分校)