We propose a new algorithm for finding an unknown number of geometric models, e.g., homographies. The problem is formalized as finding dominant model instances progressively without forming crisp point-to-model assignments. Dominant instances are found via a RANSAC-like sampling and a consolidation process driven by a model quality function considering previously proposed instances. New ones are found by clustering in the consensus space. This new formulation leads to a simple iterative algorithm with state-of-the-art accuracy while running in real-time on a number of vision problems - at least two orders of magnitude faster than the competitors on two-view motion estimation. Also, we propose a deterministic sampler reflecting the fact that real-world data tend to form spatially coherent structures. The sampler returns connected components in a progressively densified neighborhood-graph. We present a number of applications where the use of multiple geometric models improves accuracy. These include pose estimation from multiple generalized homographies; trajectory estimation of fast-moving objects; and we also propose a way of using multiple homographies in global SfM algorithms. Source code: https://github.com/danini/clustering-in-consensus-space.
Refining Pseudo Labels with Clustering Consensus over Generations for Unsupervised Object Re-identification
CUHK-SenseTime Joint Laboratory（港中大-商汤联合实验室）；The Chinese University of Hong Kong（香港中文大学）；SIAT-SenseTime Joint Lab（深圳先进院-商汤联合实验室）；Shanghai AI Laboratory（上海人工智能实验室）；Xidian University（西安电子科技大学）
Consensus Clustering With Unsupervised Representation Learning
The Ohio State University（俄亥俄州立大学）；Microsoft（微软）
深度聚类和无监督表示学习的最新进展是基于这样的想法，即输入图像的不同视图（通过数据增强技术生成）必须在表示空间中更接近，或者具有相似的聚类分配。Bootstrap Your Own Latent（BYOL）就是这样一种表示学习算法，它在线性评估协议下的ImageNet自监督图像分类中取得了最先进的结果。然而，没有探索BYOL的学习功能来执行聚类的效用。在这项工作中，我们研究了BYOL的聚类能力，并观察到使用BYOL学习的特征可能不是聚类的最佳选择。我们提出了一种新的基于一致性聚类的损失函数，并以端到端的方式使用所提出的损失进行训练BYOL，从而提高了聚类能力，并在一些流行的计算机视觉数据集上优于基于相似聚类的方法。
Consensus Clustering of Free-Viewing Gaze Data: New Insights into Human-Information Interaction
Graphics-Visualization-Computing Lab, International Institute of Information Technology Bangalore（图形-可视化-计算实验室，国际信息技术研究所班加罗尔）；PSG College of Technology（PSG理工学院）
GP2C: Geometric Projection Parameter Consensus for Joint 3D Pose and Focal Length Estimation in the Wild
Institute of Computer Graphics and Vision, Graz University of Technology, Austria（奥地利格拉茨工业大学计算机图形与视觉研究所）；Laboratoire Bordelais de Recherche en Informatique, University of Bordeaux, France（法国波尔多大学波尔多信息学实验室）
School of Artificial Intelligence, Xidian University（西安电子科技大学人工智能学院）；MoE Key Lab of Artificial Intelligence, AI Institute, Shanghai Jiao Tong University（上海交通大学人工智能研究院）
Latent Semantic Consensus For Deterministic Geometric Model Fitting
School of Electronics and Information Engineering, Tongji University（同济大学电子与信息工程学院）；School of Computer Science and Technology, Hangzhou Dianzi University（杭州电子科技大学计算机学院）；Department of Computer Science and Technology, Harbin Institute of Technology (Shenzhen)（哈尔滨工业大学（深圳）计算机科学与技术系）；Electronic Information School, Wuhan University（武汉大学电子信息学院）；Nankai International Advanced Research Institute (SHENZHEN FUTIAN), Nankai University（南开大学深圳福田国际先进研究院）；College of Computer Science, Nankai University（南开大学计算机学院）；UCAS-Terminus AI Lab, University of Chinese Academy of Sciences（中国科学院大学- Terminus人工智能实验室）
Faculty of Computer Science, University of Vienna（维也纳大学计算机科学学院）；UniVie Doctoral School Computer Science, University of Vienna（维也纳大学UniVie计算机科学博士学院）；LMU Munich（慕尼黑大学）；Munich Center for Machine Learning（慕尼黑机器学习中心）；Faculty of Philological and Cultural Studies, University of Vienna（维也纳大学语言与文化学院）
RC-GeoCP: Geometric Consensus for 4D Radar-Camera Collaborative Perception
College of Information Science and Electronic Engineering, Zhejiang University（浙江大学信息科学与电子工程学院）；School of Automotive Studies, Tongji University（同济大学汽车学院）；Thrust of Artificial Intelligence, Hong Kong University of Science and Technology（香港科技大学人工智能研究所）
