检索结果-内蒙古大学图书馆

IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

作者： Shuai, Bing Berneshawi, Andrew Li, Xinyu Modolo, Davide Tighe, Joseph Amazon Web Serv AWS Seattle WA 98109 USA

ISBN: (纸本)9781665445092

In this paper, we focus on improving online multi-object tracking (MOT). In particular, we introduce a region-based Siamese Multi-Object Tracking network, which we name SiamMOT. SiamMOT includes a motion model that estimates the instance's movement between two frames such that detected instances are associated. To explore how the motion modelling affects its tracking capability, we present two variants of Siamese tracker, one that implicitly models motion and one that models it explicitly. We carry out extensive quantitative experiments on three different MOT datasets: MOT17, TAO-person and Caltech Roadside Pedestrians, showing the importance of motion modelling for MOT and the ability of SiamMOT to substantially outperform the state-of-the-art. Finally, SiamMOT also outperforms the winners of ACM MM'20 HiEve Grand Challenge on HiEve dataset. Moreover, SiamMOT is efficient, and it runs at 17 FPS for 720P videos on a single modern GPU.

关键词： Computer vision Tracking Graphics processing units Pattern recognition Videos

来源：评论

学校读者我要写书评

暂无评论

TrackFuse: Improving tracker performance by late stage fusion 13

TrackFuse: Improving tracker performance by late stage fusio...

引用

Conference on Geospatial Informatics XIII

作者： Pradhan, Rohan Goley, Steve Welch, Austin Amazon Web Serv AWS Seattle WA 98104 USA

ISBN: (数字)9781510661653

ISBN: (纸本)9781510661646;9781510661653

Multiple object tracking (MOT) is a common computer vision problem that focuses on detecting objects and maintaining their identities through a sequence of image frames. Until now, there have been three main approaches to improve MOT performance: 1) improving the detector's quality, 2) improving the tracker's quality, or 3) creating novel approaches to jointly model detection and tracking. In this work, we argue that there is a fourth, simpler way to improve MOT performance, by fusing multiple multiple object trackers together. In this paper, we introduce a novel approach, TrackFuse, that aims to fuse the final tracks from two different models into a single output, similar to classification ensembling or weighted box fusion for object detection. The fundamental assumption of TrackFuse is that multiple trackers will fail uniquely, and similarly, multiple detectors will fail uniquely too. Thus, by fusing the output of multiple approaches to MOT, we can improve tracking performance. We test our approach on combinations of several high performing approaches to tracking and show state-of-the-art results on the MOTA metric on a held out validation set of the MOT17 dataset, compared to individual tracking models. Furthermore, we consistently show that fusing multiple object trackers provides a performance boost on multiple metrics compared to results of individual model outputs sent for fusion. Our code will be released soon.

关键词： computer vision transformers masked autoencoding remote sensing wide area motion imagery object detection

来源：评论

学校读者我要写书评

暂无评论

Large-scale Indoor Mapping with Failure Detection and Recovery in SLAM

Large-scale Indoor Mapping with Failure Detection and Recove...

引用

2024 International Conference on Intelligent Robots and Systems

作者： Rahman, Sharmin DiPietro, Robert Kedarisetti, Dharanish Kulathumani, Vinod Amazon Amazon Web Serv AWS Applicat Seattle MA 98109 USA

ISBN: (纸本)9798350377712;9798350377705

This paper addresses the failure detection and recovery problem in visual-inertial based Simultaneous Localization and Mapping (SLAM) systems for large-scale indoor environments. Camera and Inertial Measurement Unit (IMU) are popular choices for SLAM in many robotics tasks (e.g., navigation) due to their complementary sensing capabilities and low cost. However, vision has inherent challenges even in well-lit scenes, including motion blur, lack of features, or even accidental camera blockage. These failures can cause drifts to accumulate over time and can severely impact the scalability of existing solutions to large areas. To address these issues, we propose an automatic map generation service with (i) a failure detection method based on visual feature tracking quality using a health tracker which identifies and discards faulty measurements and (ii) a continuous session merging approach in SLAM. Taken together, this allows us to handle erroneous data without any manual intervention, and allows us to scale to extremely large spaces. The proposed system has been validated on benchmark datasets. Also, experimental results on multiple custom large-scale grocery stores, each between 1700 m(2) to 3700 m(2), and duration 60 to 80 minutes, are presented. Our approach shows the lowest error in all large-scale SLAM cases when compared with state-of-the-art visual-inertial SLAM packages, which often produce highly erroneous trajectories or lose track. Additionally, we provide dense 3D reconstruction with the presence of a depth camera by simply registering the point cloud from RGB-D image with respect to the SLAM generated trajectory - and the quality of the reconstruction illustrates the efficacy of our proposed method.

关键词： SLAM robotics

来源：评论

学校读者我要写书评

暂无评论

Towards Masked Autoencoding Pre-training for Wide Area Motion Imagery 13

Towards Masked Autoencoding Pre-training for Wide Area Motio...

引用

Conference on Geospatial Informatics XIII

作者： Goley, Steve Pradhan, Rohan Welch, Austin Amazon Web Serv AWS Seattle WA 98104 USA

ISBN: (数字)9781510661653

ISBN: (纸本)9781510661646;9781510661653

Transformer models are demonstrating remarkable and emergent capabilities in the natural language processing domain. These models are bounded only by the availability of large training datasets. These datasets can be tractably obtained since natural language models are pre-trained using self-supervision in the form of token masking. Papers like He et al.(1) and Cao et al.(2) have recently shown the power of this token masking technique by utilizing masked autoencoders as scalable vision learners in combination with a self-supervised pre-training technique for vision transformer models. Feichtenhofer et al.(3) extended these techniques to video, proving that masked autoencoders are scalable spatiotemporal learners as well. To our best knowledge, these techniques have only been experimented on ground-level, object-centric style imagery and video. Extending these techniques to remote or overhead imagery presents two significant problems. First, the size of objects of interest are small compared to the typical mask patch size. Second, the frames are not object centered. In this study, we explore if modern self-supervised pre-training techniques like masked auto encoding extend well to overhead wide area motion imagery (WAMI) data. We argue that modern pre-training techniques like MAE are well suited to WAMI data given the typical object size in this domain as well as the ability to leverage strong global spatial contextual information. To this end, we conduct a comprehensive exploration of different patch sizes and masking ratios on the popular WAMI dataset, WPAFB 2009.(4) We find that domain-specific adjustments to these pre-training techniques result in downstream performance improvements on computer vision tasks including object detection.

关键词： computer vision transformers masked autoencoding remote sensing wide area motion imagery object detection

来源：评论

学校读者我要写书评

暂无评论

FinLex: An effective use of word embeddings for financial lexicon generation

引用

JOURNAL OF FINANCE AND DATA SCIENCE 2022年 8卷 1-11页

作者： Das, Sanjiv R. Donini, Michele Zafar, Muhammad Bilal He, John Kenthapadi, Krishnaram Santa Clara Univ Santa Clara CA 95053 USA Amazon Inc Seattle WA 98109 USA AWS Amazon Web Serv Pforzheim Germany AWS Amazon Web Serv Seattle WA USA

We present a simple and effective methodology for the generation of lexicons (word lists) that may be used in natural language scoring applications. In particular, in the finance industry, word lists have become ubiquitous for sentiment scoring. These have been derived from dictionaries such as the Harvard Inquirer and require manual curation. Here, we present an automated approach to the curation of lexicons, which makes automatic preparation of any word list immediate. We show that our automated word lists deliver comparable performance to traditional lexicons on machine learning classification tasks. This new approach will enable finance academics and practitioners to create and deploy new word lists in addition to the few traditional ones in a facile manner. (c) 2021 The Authors. Publishing services by Elsevier B.V. on behalf of KeAi Communications Co. Ltd. This is an open access article under the CC BY-NC-ND license (http://***/licenses/by-nc-nd/4.0/).

关键词： Lexicons Embeddings Scoring Machine learning

来源：评论

学校读者我要写书评

暂无评论

Minimum k-cores and the k-core polytope

引用

NETWORKS 2022年第1期80卷 93-108页

作者： Mikesell, Derek Hicks, Illya V. Amazon Web Serv AWS Demand Modeling Grp Seattle WA 98109 USA Rice Univ Computat & Appl Math Houston TX USA

The minimum k-core problem asks for the smallest induced subgraph of minimum degree k. It has been shown that this problem is NP-hard, and thus sophisticated techniques are required to obtain good solutions and approximations. In this article, the minimum k-core problem is modeled as a binary integer program and relaxed as a linear program. Since the relaxation may yield a non-integral solution, a branch-and-cut framework is used to find an integral optimal solution. It is shown that the edge and cycle transversals of the graph give valid inequalities for the convex hull of the k-core polytope-which can be further generalized to a family of l-core transversals. Further, a heuristic for the transversal of the minimal l-cores is given with its associated valid inequality. Additionally, improved valid inequalities are generated using bounds involving the girth of the graph. Multiple heuristics are explored for finding initial bounds for the branching process utilizing the degree distribution of the graph. Finally, numerical results are given comparing the branch-and-bound, branch-and-cut, and heuristic techniques.

关键词： binary integer program branch-and-cut feedback vertex set k-core heuristics minimum k-core problem valid inequalities

来源：评论

学校读者我要写书评

暂无评论

Machine Learning-Based Profiling in Test Cheating Detection

引用

EDUCATIONAL MEASUREMENT-ISSUES AND PRACTICE 2023年第1期42卷 59-75页

作者： Meng, Huijuan Ma, Ye Amazon Web Serv AWS 410 Terry Ave N Seattle WA 98109 USA

In recent years, machine learning (ML) techniques have received more attention in detecting aberrant test-taking behaviors due to advantages when compared to traditional data forensics methods. However, defining "True Test Cheaters" is challenging-different than other fraud detection tasks such as flagging forged bank checks or credit card frauds, testing organizations are often lack of physical evidences to identify "True Test Cheaters" to train ML models. This study proposed a statistically defensible method of labeling "True Test Cheaters" in the data, demonstrated the effectiveness of using ML approaches to identify irregular statistical patterns in exam data, and established an analytical framework for evaluating and conducting real-time ML-based test data forensics. Classification accuracy and false negative/positive results are evaluated across different supervised-ML techniques. The reliability and feasibility of operationally using this approach for an IT certification exam are evaluated using real data.

关键词： machine Learning test security

来源：评论

学校读者我要写书评

暂无评论

Utility-Preserving Privacy-Enabled Speech Embeddings for Emotion Detection 24

Utility-Preserving Privacy-Enabled Speech Embeddings for Emo...

引用

Interspeech Conference

作者： Lavania, Chandrashekhar Das, Sanjiv Huang, Xin Han, Kyu J. AWS AI Labs Seattle WA 98109 USA Amazon Web Serv Seattle WA USA

Audio privacy has been undertaken using adversarial task training or adversarial models based on GANs, where the models also suppress scoring of other attributes (e.g., emotion, etc.), but embeddings still retain enough information to bypass speaker privacy. We use methods for feature importance from the explainability literature to modify embeddings from adversarial task training, providing a simple and accurate approach to generating embeddings for preserving speaker privacy while not attenuating utility for related tasks (e.g., emotion recognition). This enables better adherence with privacy regulations around biometrics and voiceprints, while retaining the usefulness of audio representation learning.

关键词： adversarial training speaker identification emotion scoring feature shuffling

来源：评论

学校读者我要写书评

暂无评论

AutoGluon-TimeSeries: AutoML for Probabilistic Time Series Forecasting 2

AutoGluon-TimeSeries: AutoML for Probabilistic Time Series F...

引用

2nd International Conference on Automated Machine Learning (AutoML)

作者： Shchur, Oleksandr Turkmen, Caner Erickson, Nick Shen, Huibin Shirkov, Alexander Hu, Tony Wang, Yuyang Amazon Web Serv Seattle WA 98109 USA AWS AI Labs Seattle WA USA

We introduce AutoGluon-TimeSeries-an open-source AutoML library for probabilistic time series forecasting.(1) Focused on ease of use and robustness, AutoGluon-TimeSeries enables users to generate accurate point and quantile forecasts with just 3 lines of Python code. Built on the design philosophy of AutoGluon, AutoGluon-TimeSeries leverages ensembles of diverse forecasting models to deliver high accuracy within a short training time. AutoGluon-TimeSeries combines both conventional statistical models, machine-learning based forecasting approaches, and ensembling techniques. In our evaluation on 29 benchmark datasets, AutoGluon-TimeSeries demonstrates strong empirical performance, outperforming a range of forecasting methods in terms of both point and quantile forecast accuracy, and often even improving upon the best-in-hindsight combination of prior methods.

关键词：

来源：评论

学校读者我要写书评

暂无评论

XPM Efficiency Versus Symbol Rate

引用

JOURNAL OF LIGHTWAVE TECHNOLOGY 2022年第9期40卷 2850-2861页

作者： Lowery, Arthur J. Du, Liang B. Y. Monash Univ Dept Elect & Comp Syst Engn Electrophoton Lab Melbourne Vic 3800 Australia Amazon Web Serv AWS San Carlos CA USA

Cross-Phase Modulation (XPM) limits the information capacity of optical fibers, due to the power fluctuations in one channel affecting the phases of the other channels. A decade ago, we showed that splitting a channel into multiple subcarriers could reduce the effect of XPM on multi-channel systems, due to a reduction in the intensity fluctuations at frequencies where walk-off is low. As commercial systems now take advantage of this effect, it is appropriate to give a more detailed and quantitative explanation of it;in particular to show that the optimal subcarrier rate scales with the square-root of the chirp rate of the symbols, which depends only on the accumulated dispersion in the fiber. Simulations show that the optimal symbol rate is when the time-spread of the symbols due to dispersion is between 2.5x to 6.5x of their transmitted (central pulse) widths. This paper presents simple arguments backed by simulations to identify the key parameters that determine the optimal subcarrier rate. In particular, we show that there is an optimal balance between the intrinsic intensity fluctuations of transmitted symbols (which are lower-frequency for lower symbol rates) and the intensity fluctuations due to chirp (which are lower-frequency for higher symbol rates). We also discuss why the traditional XPM-efficiency equation is not appropriate when data is transmitted, because it assumes the XPM contributions of all spans in an amplified link are equal, rather than increasing with distance.

关键词： Modulation Fluctuations Optical fiber dispersion Mathematical models Chirp Frequency modulation Optical fiber amplifiers Chirp cross-phase modulation dispersion enhanced Gaussian noise fiber nonlinearity Nyquist sub-carriers OFDM optical fiber communications symbol rate optimization XPM efficiency

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：