检索结果-内蒙古大学图书馆

Curvature-guided adaptive T-spline surface fitting

COMPUTER-AIDED DESIGN 2013年第8-9期45卷 1095-1107页

作者： Wang, Yimin Zheng, Jianmin Shanghai Univ Sch Comp Engn & Sci Shanghai 200444 Peoples R China Nanyang Technol Univ Sch Comp Engn Singapore 639798 Singapore

The problem of fitting spline surfaces to triangular mesh models is of importance in computer-aided design. Many fitting algorithms have been developed. This paper proposes several novel plug-and-play components or strategies: the use of T-splines for fitting, a curvature-guided strategy, faithful re-parameterization and initial spline knot re-placement, which can be used to enhance fitting algorithms. We also present an adaptive T-spline fitting algorithm integrating these components and strategies. Extensive experiments have been conducted to demonstrate these components. Our fitting algorithm can generate spline surfaces that well respect the geometrical features of input mesh models and have a more compact representation. (C) 2013 Elsevier Ltd. All rights reserved.

关键词： Surface fitting T-splines Curvature re-parameterization Knot re-placement

来源：评论

学校读者我要写书评

暂无评论

Expanding and refining Hybrid Compressors for Efficient Object re-Identification

引用

IEEE TRANSACTIONS ON IMAGE PROCESSING 2024年 33卷 3793-3808页

作者： Xie, Yi Wu, Hanxiao Zhu, Jianqing Zeng, Huanqiang Zhang, Jing Huaqiao Univ Coll Engn Quanzhou 362021 Peoples R China South China Univ Technol Sch Future Technol Guangzhou 510000 Peoples R China Huaqiao Univ Coll Informat Sci & Engn Xiamen 361021 Peoples R China Wuhan Univ Technol Sch Comp Sci & Artificial Intelligence Wuhan 430070 Peoples R China Univ Sydney Sch Comp Sci Sydney NSW 2006 Australia

recent object re-identification (re-ID) methods gain high efficiency via lightweight student models trained by knowledge distillation (KD). However, the huge architectural difference between lightweight students and heavy teachers causes students to have difficulties in receiving and understanding teachers' knowledge, thus losing certain accuracy. To this end, we propose a refiner-expander-refiner (reR) structure to enlarge a student's representational capacity and prune the student's complexity. The expander is a multi-branch convolutional layer to expand the student's representational capacity to understand a teacher's knowledge comprehensively, which does not require any feature-dimensional adapter to avoid knowledge distortions. The two refiners are 1 x 1 convolutional layers to prune the input and output channels of the expander. In addition, in order to alleviate the competition accuracy-related and pruning-related gradients, we design a common consensus gradient resetting (CCGR) method, which discards unimportant channels according to the intersection of each sample's unimportant channel judgment. Finally, the trained reR can be simplified into a slim convolutional layer via re-parameterization to speed up inference. As a result, we propose an expanding and refining hybrid compressing (ERHC) method. Extensive experiments show that our ERHC has superior inference speed and accuracy, e.g., on the VeRi-776 dataset, given the resNet101 as a teacher, ERHC saves 75.33% model parameters (MP) and 74.29% floating-point of operations (FLOPs) without sacrificing accuracy.

关键词： Object re-identification knowledge distillation network pruning re-parameterization

来源：评论

学校读者我要写书评

暂无评论

Patch-based image reconstruction for PET using prior-image derived dictionaries

引用

PHYSICS IN MEDICINE AND BIOLOGY 2016年第18期61卷 6833-6855页

作者： Tahaei, Marzieh S. reader, Andrew J. McGill Univ Montreal Neurol Inst McConnell Brain Imaging Ctr Montreal PQ Canada Kings Coll London St Thomas Hosp Dept Biomed Engn Div Imaging Sci & Biomed Engn London England

In PET image reconstruction, regularization is often needed to reduce the noise in the resulting images. Patch-based image processing techniques have recently been successfully used for regularization in medical image reconstruction through a penalized likelihood framework. re-parameterization within reconstruction is another powerful regularization technique in which the object in the scanner is re-parameterized using coefficients for spatially-extensive basis vectors. In this work, a method for extracting patch-based basis vectors from the subject's MR image is proposed. The coefficients for these basis vectors are then estimated using the conventional MLEM algorithm. Furthermore, using the alternating direction method of multipliers, an algorithm for optimizing the Poisson log-likelihood while imposing sparsity on the parameters is also proposed. This novel method is then utilized to find sparse coefficients for the patch-based basis vectors extracted from the MR image. The results indicate the superiority of the proposed methods to patch-based regularization using the penalized likelihood framework.

关键词： positron emission tomography image reconstruction anatomical prior hybrid PET-MR patch representation sparsity re-parameterization

来源：评论

学校读者我要写书评

暂无评论

TMS: Temporal multi-scale in time-delay neural network for speaker verification

引用

APPLIED INTELLIGENCE 2023年第22期53卷 26497-26517页

作者： Zhang, Ruiteng Wei, Jianguo Lu, Xugang Lu, Wenhuan Jin, Di Zhang, Lin Xu, Junhai Dang, Jianwu Tianjin Univ Coll Intelligence & Comp Tianjin Peoples R China Qinghai Nationalities Univ Comp Coll Xining Qinghai Peoples R China Natl Inst Informat & Commun Technol Kyoto Japan Natl Inst Informat Tokyo Japan

The speaker encoder is an important front-end module that explores discriminative speaker features for many speech applications requiring speaker information. Current speaker encoders aggregate multi-scale features from utterances using multi-branch network architectures. However, naively adding many branches through a fully convolutional operation cannot efficiently improve its capability to capture multi-scale features due to the problem of rapid increase of model parameters and computational complexity. Therefore, in current network architectures, only a few branches corresponding to a limited number of temporal scales are designed for capturing speaker features. To address this problem, this paper proposes an effective temporal multi-scale (TMS) model where multi-scale branches could be efficiently designed in a speaker encoder while negligibly increasing computational costs. The TMS model is based on a time-delay neural network (TDNN), where the network architecture is separated into channel-modeling and temporal multi-branch modeling operators. In the TMS model, adding temporal multi-scale elements in the temporal multi-branch operator only slightly increases the model's parameters, thus saving more of the computational budget to add branches with large temporal scales. After model training, we further develop a systemic re-parameterization method to convert the multi-branch network topology into a single-path-based topology to increase the inference *** conducted automatic speaker verification (ASV) experiments under in-domain (VoxCeleb) and out-of-domain (CNCeleb) conditions to investigate the proposed TMS model's *** results show that the TMS-method-based model outperformed state-of-the-art ASV models (e.g., ECAPA-TDNN) and improved robustness. Moreover, the proposed model achieved a 29%-46% increase in the inference speed compared to ECAPA-TDNN.

关键词： Speaker verification Temporal multi-scale feature Effective backbone re-parameterization

来源：评论

学校读者我要写书评

暂无评论

Lightweight detection model for coal gangue identification based on improved YOLOv5s

引用

JOURNAL OF reAL-TIME IMAGE PROCESSING 2024年第4期21卷 1-11页

作者： Shang, Deyong Lv, Zhibin Gao, Zehua Li, Yuntao China Univ Min & Technol Beijing Sch Mech Elect & Informat Engn Beijing Peoples R China China Univ Min & Technol Beijing Intelligent Min & Robot Res Inst Beijing Peoples R China

Focusing on the issues of complex models, high computational cost, and low identification speed of existing coal gangue image identification object detection algorithms, an optimized YOLOv5s lightweight detection model for coal gangue is proposed. Using ShuffleNetV2 as the backbone network, a convolution pooling module is used at the input end instead of the original convolution module. Combining the re-parameterization idea of repVGG and introducing depthwise separable convolution, a neck feature fusion network is constructed. And using the WIoU function as the loss function. The experimental findings indicate that the improved model maintains the same accuracy, the number of parameters is only 5.1% of the original, the computational effort is reduced to 6.3 % of the original, and the identification speed is improved by 30.9% on GPU and 4 times on CPU. This method significantly reduces model complexity and improves detection speed while maintaining detection accuracy.

关键词： YOLOv5s Lightweight Coal gangue identification re-parameterization Loss function

来源：评论

学校读者我要写书评

暂无评论

re-Parameterized real-Time Stereo Matching Network Based on Mixed Cost Volumes Toward Autonomous Driving

引用

IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS 2023年第12期24卷 14914-14926页

作者： Yao, Bowen Wei, Wei Huang, Jinhao Liang, Bifa Li, Jun Guangzhou Univ Sch Elect & Commun Engn Guangzhou 510006 Peoples R China

3D perception is an essential capability of autonomous vehicles. Most state-of-the-art stereo matching networks pursue higher prediction accuracy at the cost of the inference speed. However, high demand on computational resource pushes the cost of hardware, hindering practical applications of stereo matching. In this paper, we propose RMCNet, a novel re-parameterized coarse-to-fine network for stereo matching. RMCNet achieves real time on edge devices while outputs accurate disparity maps. To reduce the computing complexity of 3D convolution in cost aggregation, we propose Mixed Cost Volumes that consist of 4D and 3D cost volumes, enabling efficient optimization of the initial disparity map for real-time inference. An Efficient Feature Matching module is proposed to provide absolute disparity candidates for 3D cost volume. To increase the reliability of 4D and 3D cost volumes, we propose a simple but efficient 3D cost aggregation module and a pyramid 2D cost aggregation module for the aggregation of Mixed Cost Volumes. Moreover, we propose a re-parameterized Channel Sensing module and a re-parameterized Disparity-Aware module to replace 2D and 3D convolutional blocks with residual connections. Being evaluated on KITTI 2012 and KITTI 2015, the comprehensive performance of the proposed network is improved by the re-parameterization approach. Outstanding trade-off between accuracy and speed on edge devices over existing SOTA works is achieved. With fast average inference speed on NVIDIA Jetson Nano of 33.74 FPS, the accuracy in terms of 3px-all on KITTI 2012 test set and D1-all on KITTI 2015 test set are 3.01% and 3.41%, respectively.

关键词： Stereo matching autonomous driving cost construction re-parameterization edge AI

来源：评论

学校读者我要写书评

暂无评论

Multi Path real-time Semantic Segmentation Network in Road Scenarios

引用

INTERNATIONAL JOURNAL OF INTELLIGENT TRANSPORTATION SYSTEMS reSEARCH 2024年第1期23卷 47-60页

作者： Pengfei, Gao Xiaolong, Tian Cuihong, Liu Chenfei, Yang Zhengzhou Railway Vocat & Tech Coll Zhengzhou 450000 Henan Peoples R China Henan Rail Transit Intelligent Safety Engn Technol Zhengzhou 450000 Henan Peoples R China Zhengzhou Key Lab Adv Funct Mat Zhengzhou 450000 Henan Peoples R China

Semantic segmentation is a critical task in computer vision. Existing methods often struggle to balance accuracy and computational efficiency when processing high-resolution images, limiting their application scenarios. To address these limitations, we introduce repMPSeg, a novel re-parameterization-based multi-path real-time semantic segmentation network. repMPSeg improves upon traditional dual-branch architectures. It features two independent yet interconnected branches: a high-resolution branch for detailed feature capture and a low-resolution branch for global semantic information extraction. By employing re-parameterization techniques, the basic convolutional blocks are optimized to enhance feature capture and local context information. During training, parallel convolution structures are utilized, which are then streamlined into a single kernel during inference to maintain performance while reducing computational complexity. The high-resolution branch leverages sub-pixel sampling and 1x\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\times $$\end{document}1 convolutions to improve the receptive field and minimize computational load, while the low-resolution branch uses 4x\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\times $$\end{document} downsampling to enhance semantic information extraction. Features from both branches are fused through the re-parameterization-based Hybrid Downsampling Module (repHDM), which aligns and combines features effectively. Our experiments on the Cityscapes and CamVid datasets demonstrate that repMPSeg achieves a balance between speed and accuracy, outperforming state-of-

关键词： real-time semantic segmentation re-parameterization High-resolution images Dual-branch architecture

来源：评论

学校读者我要写书评

暂无评论

Multi-GNSS precise point positioning (MGPPP) using raw observations

引用

JOURNAL OF GEODESY 2017年第3期91卷 253-268页

作者： Liu, Teng Yuan, Yunbin Zhang, Baocheng Wang, Ningbo Tan, Bingfeng Chen, Yongchang Chinese Acad Sci Inst Geodesy & Geophys State Key Lab Geodesy & Earths Dynam 340 Xudong Rd Wuhan 430077 Peoples R China Univ Chinese Acad Sci 19 Yuquan Rd Beijing 100049 Peoples R China

A joint-processing model for multi-GNSS (GPS, GLONASS, BDS and GALILEO) precise point positioning (PPP) is proposed, in which raw code and phase observations are used. In the proposed model, inter-system biases (ISBs) and GLONASS code inter-frequency biases (IFBs) are carefully considered, among which GLONASS code IFBs are modeled as a linear function of frequency numbers. To get the full rank function model, the unknowns are re-parameterized and the estimable slant ionospheric delays and ISBs/IFBs are derived and estimated simultaneously. One month of data in April, 2015 from 32 stations of the International GNSS Service (IGS) Multi-GNSS Experiment (MGEX) tracking network have been used to validate the proposed model. Preliminary results show that RMS values of the positioning errors (with respect to external double-difference solutions) for static/kinematic solutions (four systems) are 6.2 mm/2.1 cm (north), 6.0 mm/2.2 cm (east) and 9.3 mm/4.9 cm (up). One-day stabilities of the estimated ISBs described by STD values are 0.36 and 0.38 ns, for GLONASS and BDS, respectively. Significant ISB jumps are identified between adjacent days for all stations, which are caused by the different satellite clock datums in different days and for different systems. Unlike ISBs, the estimated GLONASS code IFBs are quite stable for all stations, with an average STD of 0.04 ns over a month. Single-difference experiment of short baseline shows that PPP ionospheric delays are more precise than traditional leveling ionospheric delays.

关键词： Multi-GNSS precise point positioning Raw observations re-parameterization Inter-system bias Inter-frequency bias PPP ionospheric delays

来源：评论

学校读者我要写书评

暂无评论

Estimating m-regimes STAR-GARCH model using QMLE with parameter transformation

引用

MATHEMATICS AND COMPUTERS IN SIMULATION 2011年第7期81卷 1385-1396页

作者： Chan, Felix Theoharakis, Billy Curtin Univ Technol Sch Econ & Finance Bentley WA 6102 Australia

It is well known in the literature that obtaining the parameter estimates for the Smooth Transition Autoregressive-Generalized Autoregressive Conditional Heteroskedasticity (STAR-GARCH) can be problematic due to computational difficulties. Conventional optimization algorithms do not seem to perform well in locating the global optimum of the associated likelihood function. This makes Quasi-Maximum Likelihood Estimator (QMLE) difficult to obtain for STAR-GARCH models in practice. Curiously, there has been very little research investigating the cause of the numerical difficulties in obtaining the parameter estimates for STAR-GARCH using QMLE. The aim of the paper is to investigate the nature of the numerical difficulties using Monte Carlo Simulation. By examining the surface of the log-likelihood function based on simulated data, the results provide several insights into the difficulties in obtaining QMLE for STAR-GARCH models. Based on the findings, the paper also proposes a simple transformation on the parameters to alleviate these difficulties. Monte Carlo simulation results show promising signs for the proposed transform. The asymptotic and robust variance covariance matrices of the original parameter estimates are derived as a function of the transformed parameter estimates, which greatly facilitates inferences on the original parameters. (c) 2010 IMACS. Published by Elsevier B.V. All rights reserved.

关键词： STAR GARCH Monte Carlo simulation re-parameterization

来源：评论

学校读者我要写书评

暂无评论

real Spike: Learning real-Valued Spikes for Spiking Neural Networks 1

引用

17th European Conference on Computer Vision (ECCV)

作者： Guo, Yufei Zhang, Liwen Chen, Yuanpei Tong, Xinyi Liu, Xiaode Wang, YingLei Huang, Xuhui Ma, Zhe Intelligent Sci & Technol Acad CASIC Beijing 100854 Peoples R China

ISBN: (数字)9783031197758

ISBN: (纸本)9783031197741;9783031197758

Brain-inspired spiking neural networks (SNNs) have recently drawn more and more attention due to their event-driven and energy-efficient characteristics. The integration of storage and computation paradigm on neuromorphic hardwares makes SNNs much different from Deep Neural Networks (DNNs). In this paper, we argue that SNNs may not benefit from the weight-sharing mechanism, which can effectively reduce parameters and improve inference efficiency in DNNs, in some hardwares, and assume that an SNN with unshared convolution kernels could perform better. Motivated by this assumption, a training-inference decoupling method for SNNs named as real Spike is proposed, which not only enjoys both unshared convolution kernels and binary spikes in inference-time but also maintains both shared convolution kernels and real-valued Spikes during training. This decoupling mechanism of SNN is realized by a re-parameterization technique. Furthermore, based on the training-inference-decoupled idea, a series of different forms for implementing real Spike on different levels are presented, which also enjoy shared convolutions in the inference and are friendly to both neuromorphic and non-neuromorphic hardware platforms. A theoretical proof is given to clarify that the real Spike-based SNN network is superior to its vanilla counterpart. Experimental results show that all different real Spike versions can consistently improve the SNN performance. Moreover, the proposed method outperforms the state-of-the-art models on both non-spiking static and neuromorphic datasets.

关键词： Spiking neural network real spike Binary spike Training-inference-decoupled re-parameterization

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：