3,287 research outputs found
Unsupervised Point Cloud Representation Learning by Clustering and Neural Rendering
No full textData augmentation has contributed to the rapid advancement of unsupervised learning on 3D point clouds. However, we argue that data augmentation is not ideal, as it requires a careful application-dependent selection of the types of augmentations to be performed, thus potentially biasing the information learned by the network during self-training. Moreover, several unsupervised methods only focus on uni-modal information, thus potentially introducing challenges in the case of sparse and textureless point clouds. To address these issues, we propose an augmentation-free unsupervised approach for point clouds, named CluRender, to learn transferable point-level features by leveraging uni-modal information for soft clustering andcross-modal information for neural rendering. Soft clustering enables self-training through a pseudo-label prediction task, where the affiliation of points to their clusters is used as a proxy under the constraint that these pseudo-labels divide the point cloud into approximate equal partitions. This allows us to formulate a clustering loss to minimize the standard cross-entropy betweenpseudoandpredictedlabels.Neuralrenderinggenerates photorealistic renderings from various viewpoints to transfer photometric cues from 2D images to the features. The consistency between rendered and real images is then measured to form a fitting loss, combined with the cross-entropy loss to self-train networks. Experiments on downstream applications, including 3D object detection, semantic segmentation, classification, part segmentation, and few-shot learning, demonstrate the effectiveness of our framework in outperforming state-of-the-art techniques
Cross-source point cloud registration: Challenges, progress and prospects
No full textThe emerging topic of cross-source point cloud (CSPC) registration has attracted increasing attention with the fast development background of 3D sensor technologies. Different from the conventional same-source point clouds that focus on data from same kind of 3D sensor (e.g., Kinect), CSPCs come from different kinds of 3D sensors (e.g., Kinect and LiDAR). CSPC registration generalizes the requirement of data acquisition from same-source to different sources, which leads to generalized applications and combines the advantages of multiple sensors. In this paper, we provide a systematic review on CSPC registration. We first present the characteristics of CSPC, and then summarize the key challenges in this research area, followed by the corresponding research progress consisting of the most recent and representative developments on this topic. Finally, we discuss the important research directions in this vibrant area and explain the role in several application fields
The Consistent Transmission Laws Under the Consistent Intervention Policy
No full textCOVID-19 has been spreading globally from 2019 to 2022, prompting many countries to establish consistent, timely, and strict intervention policies during this period. However, the existing analysis of central epidemiological parameters (CPPs) falls short in exploring the consistency of transmission laws implied by these policies. To resolve this limitation, we propose a consistent transmission law inference framework. This framework first provides a coarse-to-fine detection method to extract outbreaks and devises an autonomous inference sliding window (ASW) algorithm to estimate the infected-case-dependent CPPs of these outbreaks. Then, we reveal consistent transmission laws through a differential analysis of the estimated CPPs. Focusing on the transmission of COVID-19 in China, a typical country with a consistent intervention policy, our model reveals remarkably consistent outbreak laws. In short, the policy controls the effective growth rate almost to zero (to a 1E-03 scale) within three days since the outbreak started. Specifically: 1) the policy was very effective at the beginning of the outbreak, leading to the exposure rate hitting the bottom on one day between the second and eleventh days and then keeping it at a plateau; 2) under large-scale nucleic acid testing and contact tracking, the incubation rate decreased and reached a plateau mainly on the third day; and 3) the recovery rate fluctuated extremely little on a 1E-03 scale, showing no significant breakthrough in COVID-19 treatment that helped the policy work. Our method can be readily adapted to other countries and SEIR epidemic
Geometrically-driven Aggregation for Zero-shot 3D Point Cloud Understanding
No full textZero-shot 3D point cloud understanding can be achieved via 2D Vision-Language Models (VLMs). Existing strategies directly map VLM representations from 2D pixels of rendered or captured views to 3D points, overlooking the inherent and expressible point cloud geometric structure. Geometrically similar or close regions can be exploited for bolstering point cloud understanding as they are likely to share semantic information. To this end, we introduce the first training-free aggregation technique that leverages the point cloud's 3D geometric structure to improve the quality of the transferred VLM representations. Our approach operates iteratively, performing local-to-global aggregation based on geometric and semantic point-level reasoning. We benchmark our approach on three downstream tasks, including classification, part segmentation, and semantic segmentation, with a variety of datasets representing both synthetic/real-world, and indoor/outdoor scenarios. Our approach achieves new state-of-the-art results in all benchmarks. Code and dataset are available at https://luigiriz.github.io/geoze-website
TVEG: Model Selection of the Time-Varying Exponential Family Distributions Graphical Models
No full textThe undirected graphical model, a popular class of statistical model, offers a way to describe and explain the relationships among a set of variables. However, it remains a challenge to choose a certain graphical model to explain the relationships of variables adequately, especially when the relationships of variables are rewiring over time. This paper proposes the Time-Varying Exponential Family Distributions Graphical (TVEG) models, with time-varying structures and exponential family node-wise conditional distributions. TVEG models extend the scope of available graph models and can be applied to time-varying and exponential family distribution observation data in reality. We propose the Temporally Smoothed L1-regularized exponential family graphical estimator (TSLEG), an estimator to infer the structure of TVEG from observations. We derive sufficient conditions for the TSLEG to recover the block partition and sparse pattern with high probability. We derive a message-passing optimization method to solve the TSLEG for time-varying Ising, Gaussian, exponential, and Poisson graphs based on the ADMM. The synthetic network simulations corroborate the theoretical analysis. Analysing of real data of stocks and the US Senate by the time-varying exponential model and Poisson model indicates the effectiveness and practicality of TVEG models
Multimodal Fusion SLAM with Fourier Attention
No full textVisual SLAM is particularly challenging in environments affected by noise, varying lighting conditions, and darkness. Learning-based optical flow algorithms can leverage multiple modalities to address these challenges, but traditional optical flow-based visual SLAM approaches often require significant computational resources. To overcome this limitation, we propose FMF-SLAM, an efficient multimodal fusion SLAM method that utilizes fast Fourier transform (FFT) to enhance the algorithm efficiency. Specifically, we introduce a novel Fourier-based self-attention and cross-attention mechanism to extract features from RGB and depth signals. We further enhance the interaction of multimodal features by incorporating multi-scale knowledge distillation across modalities. We also demonstrate the practical feasibility of FMF-SLAM in real-world scenarios with real time performance by integrating it with a security robot by fusing with a global positioning module GNSS-RTK and global Bundle Adjustment. Our approach is validated using video sequences from TUM, TartanAir, and our real-world datasets, showcasing state-of-the-art performance under noisy, varying lighting, and dark conditions
New Roads for Patron-Driven E-books:Collection Development and Technical Services Implications of a Patron-Driven Acquisitions Pilot at Rutgers
No full textCollection development librarians have long struggled to meet user demands for new titles. Too often, required resources are not purchased, while some purchased resources do not circulate. E-books selected through patron-driven plans are a solution but present new challenges for both selectors and catalogers. Radical changes to traditional technical services workflows are required, and selectors must modify the selection process to give more choice to the user. Rutgers University librarians have adopted an innovative new technical services workflow and collection-development model to manage a successful, patron-driven acquisitions project for e-books in the fields of math and computer science.This is an Accepted Manuscript of an article published by Taylor & Francis Group in Journal of Electronic Resources Librarianship on 13/12/2011, available online at: http://www.tandfonline.com/10.1080/1941126X.2011.627043
Attentive Multimodal Fusion for Optical and Scene Flow
No full textThis letter presents an investigation into the estimation of optical and scene flow using RGBD information in scenarios where the RGB modality is affected by noise or captured in dark environments. Existing methods typically rely solely on RGB images or fuse the modalities at later stages, which can result in lower accuracy when the RGB information is unreliable. To address this issue, we propose a novel deep neural network approach named FusionRAFT, which enables early-stage information fusion between sensor modalities (RGB and depth). Our approach incorporates self- and cross-attention layers at different network levels to construct informative features that leverage the strengths of both modalities. Through comparative experiments, we demonstrate that our approach outperforms recent methods in terms of performance on the synthetic dataset FlyingThings3D, as well as the generalization on the real-world dataset KITTI. We illustrate that our approach exhibits improved robustness in the presence of noise and low-lighting conditions that affect the RGB images
The politics of fashion: perceptions of power in female clothing and ornamentation as reflected in the sixteenth-century Chinese novel Jin Ping Mei
No full textThis thesis examines issues of female power and influence in sixteenth-century China focusing on how women and their roles were perceived in the changing social environment of the mid-late Ming dynasty. Using aspects of a New Historicist approach, information from contemporary literary and historical sources are analysed alongside each other. With its emphasis on the lives of women and preoccupation with the description of material objects, the late Ming novel Jin Ping Mei forms an important element in the thesis. China in the sixteenth century saw expanding urbanisation, the emergence of a new wealthy merchant class, increasing visibility of women and a questioning of traditional morality. Fashion consciousness, as one of the most conspicuous aspects of the new material culture, is a possible indicator of these trends. Traditional Western theories contend that fashion began in the particular context of Renaissance Europe. However, this study argues that a similar fashion awareness existed in China too, and was manifested in a competitive striving for social status, in this case specifically among women. In contrast to previous studies which downplayed the impact women had on defining traditional Chinese culture, this thesis demonstrates how women and their sartorial choices began to redefine the boundaries of material culture, influencing literati discourse which, in turn, re- influenced female behaviour
- …
