检索结果-内蒙古大学图书馆

A robust framework combined saliency detection and image recognition for garbage classification

WASTE MANAGEMENT 2022年 140卷 193-203页

作者： Qin, Jiongming Wang, Cong Ran, Xu Yang, Shaohua Chen, Bin Southwest Univ Coll Elect & Informat Engn Chongqing Key Lab Nonlinear Circuits & Intelligen Chongqing 400715 Peoples R China

Using deep learning to solve garbage classification has become a hot topic in computer version. The most widely used garbage dataset Trashnet only has garbage images with a white board as background. Previous studies based on Trashnet focus on using different networks to achieve a higher classification accuracy without considering the complex backgrounds which might encounter in practical applications. To solve this problem, we propose a framework that combines saliency detection and image classification to improve the generalization performance and robustness. A saliency network Salinet is adopted to obtain the garbage target area. Then, a smallest rectangle containing this area is created and used to segment the garbage. A classification network Inception V3 is used to identify the segmented garbage image. Images of the original Trashnet are fused with complex backgrounds of the other saliency detection datasets. The fused and original Trashnet are used together for training to improve the robustness to noises and complex backgrounds. Compared with the image classification networks and classic target detection algorithms, the proposed framework improves the accuracy of 0.50% -15.79% on the testing sets fused with complex backgrounds. In addition, the proposed framework achieves the best performance with a gain of 4.80% in accuracy on the collected actual dataset. The comparisons prove that our framework is more robust to garbage classification in complex backgrounds. This method can be applied to smart trash cans to achieve automatic garbage classification.

关键词： Saliency detection Image segmentation Garbage classification Data fusion

来源：评论

学校读者我要写书评

暂无评论

Observer and Command-Filter-Based Adaptive Neural Network Control Algorithms for nonlinear Multi-agent Systems with Input Delay

引用

COGNITIVE COMPUTATION 2022年第2期14卷 814-827页

作者： Ma, Lidan Wang, Xin Zhou, Yuhao Southwest Univ Coll Elect & Informat Engn Chongqing Key Lab Nonlinear Circuits & Intelligen Chongqing 400044 Peoples R China

Over the last decades, many researchers have investigated the distributed adaptive consensus tacking control algorithm of multi-agent systems (MASs). Nevertheless, the existing works involving the command-filter-based adaptive consensus problem for nonlinear multi-agent systems subjected to the unmeasurable states are relatively few. Besides that, the immeasurable states and the input delay will bring few challenging in dealing with the consensus problem for MASs. (1) The radial basis function neural networks (RBF NNs) are utilized to approximate the unknown nonlinear functions and the NN-based observer is established to copy with the unmeasurable states. (2) The backstepping design method of distributed adaptive consensus control is put forward on basis of the command filtering method, which overcomes the complexity explosion problem and eliminates errors by introducing compensation signals. (3) The Pade approximation approach is served to remove the obstacle originating from the input delay. This paper addresses the observer and command-filter-based adaptive tracking control problem for nonlinear multi-agent systems with the unmeasurable states and input delay under the directed graph. The Lyapunov stability theory is utilized to prove that the proposed approach can ensure that all signals in the closed-loop system reach cooperatively semi-globally uniformly ultimately bounded (CSUUB). The simulation result is presented, and it further manifests that the effectiveness of this scheme.

关键词： Tracking control nonlinear multi-agent systems Unmeasurable states Input delay Command filtering

来源：评论

学校读者我要写书评

暂无评论

LSPP: A Novel Path Planning Algorithm Based on Perceiving Line Segment Feature

引用

IEEE SENSORS JOURNAL 2022年第1期22卷 720-731页

作者： Zhang, Yifei Wang, Shiyuan Southwest Univ Coll Elect & Informat Engn Chongqing Key Lab Nonlinear Circuits & Intelligen Chongqing 400715 Peoples R China

Path planning in the automatic driving system generally uses point features to evaluate the influence of obstacles on the robot, leading to missing information or increasing computational burden. This paper proposes a novel path planning with line segment (LSPP) algorithm from the viewpoint of line segment, which is easy to be obtained directly from sensors. In LSPP, an artificial potential field (APF) based on the azimuth and distance (ADAPF) from obstacles to the robot is first constructed for implementation of integral calculation. Unlike the traditional APF which only considers the nearest distance from obstacles to the robot, ADAPF can effectively assess the influence of boundary points on the robot by fully utilizing the information of obstacles. Based on ADAPF, boundary points are integrated into lines as edges of the drivable area in a specialized coordinate system. Then, a multi-loss function is designed to guarantee that the robot is far away from obstacles to the optimized path, and thus solved by a modified Dogleg (MDL) algorithm combined with symmetric rank-1 (SR1) method based on adaptive initial trust-region radii. Finally, a smooth and safe path is provided in LSPP with the requirements of speed and quality being satisfied simultaneously. Real-world experimental results validate the robustness and the effectiveness of LSPP.

关键词： Path planning Optimization Robots Heuristic algorithms Roads Sensors Robot kinematics Path optimization path planning automatic driving mobile robot

来源：评论

学校读者我要写书评

暂无评论

High-order consensus graph learning for incomplete multi-view clustering

引用

APPLIED INTELLIGENCE 2025年第6期55卷 1-25页

作者： Guo, Wei Che, Hangjun Leung, Man-Fai Southwest Univ Coll Elect & Informat Engn Chongqing 400715 Peoples R China Chongqing Key Lab Nonlinear Circuits & Intelligent Chongqing 400715 Peoples R China Anglia Ruskin Univ Fac Sci & Engn Cambridge CB1 1PT England

Incomplete Multi-View Clustering (IMVC) aims to partition data with missing samples into distinct groups. However, most IMVC methods rarely consider the high-order neighborhood information of samples, which represents complex underlying interactions, and often neglect the weights of different views. To address these issues, we propose a High-order Consensus Graph Learning (HoCGL) model. Specifically, we integrate a reconstruction term to recover the incomplete multi-view data. High-order proximity matrices are constructed, and the self-representation similarity matrices and multiple high-order proximity matrices are learned mutually, allowing the similarity matrices to incorporate complex high-order information. Finally, the consensus graph representation is derived from the similarity matrices through a self-weighted strategy. An efficient algorithm is designed to solve the proposed model. The excellent clustering performance of the proposed model is validated by comparing it with eight state-of-the-art models across nine datasets.

关键词： Multi-view High-order Similarity matrices Self-weighted strategy

来源：评论

学校读者我要写书评

暂无评论

Widely Linear Maximum Complex Correntropy Criterion Affine projection Algorithm and Its Performance Analysis

引用

IEEE TRANSACTIONS ON SIGNAL proCESSING 2022年 70卷 3540-3550页

作者： Qiu, Chen Qian, Guobing Wang, Shiyuan Southwest Univ Coll Elect & Informat Engn Chongqing Key Lab Nonlinear Circuits & Intelligen Chongqing 400715 Peoples R China

Recently, affine projection algorithm has been extensively studied in the Gaussian noise environment. However, the performance of affine projection algorithm will deteriorate rapidly in the presence of impulsive noise and other non-Gaussian noise. To address this issue, this paper proposes a novel affine projection algorithm based on the complex Gaussian kernel function, called widely linear maximum complex correntropy criterion affine projection algorithm (WL-MCCC-APA). With the maximum complex correntropy criterion (MCCC), the robustness of affine projection algorithm is improved significantly in the non-Gaussian noise environment, especially in the presence of impulsive noise. Meanwhile, a variable step size WL-MCCC-APA (VSS-WL-MCCC-APA) method based on fixed-point theory is proposed to further balance the convergence rate and steady-state misalignment of the algorithm. Finally, the step size boundary analysis and steady-state analysis are presented for theoretical analysis. Simulation results verify the correctness of obtained theoretical results and the superiority of the proposed WL-MCCC-APA and VSS-WL -MCCC-APA.

关键词： Signal processing algorithms projection algorithms Optimized production technology Kernel Filtering algorithms Cost function Convergence Widely linear affine projection maximum complex correntropy criterion variable step size

来源：评论

学校读者我要写书评

暂无评论

Fixed-Time Synchronization of Neural Networks with Parameter Uncertainties via Quantized Intermittent Control

引用

NEURAL proCESSING LETTERS 2022年第3期54卷 2303-2318页

作者： Yang, Wenqiang Huang, Junjian Wang, Xin Southwest Univ Coll Elect & Informat Engn Chongqing Key Lab Nonlinear Circuits & Intelligen Chongqing 400715 Peoples R China

In this paper, we study the fixed-time synchronization (FIXTS) of neural networks (NNs) with parameter uncertainties via quantized intermittent control. Based on the intermittent control strategy and quantitative control theory, sufficient conditions are established to achieve synchronization of NNs and the synchronization time can be estimated. In addition, this paper also considers the synchronization of NNs under different situations. Finally, a simulation example is given to verify the correctness of the proposed theory.

关键词： Fixed-time synchronization Quantized intermittent controller Neural networks Parameter uncertainties

来源：评论

学校读者我要写书评

暂无评论

A Multi-Objective Particle Swarm Optimization Pruning on Photonic Neural Networks

引用

JOURNAL OF LIGHTWAVE TECHNOLOGY 2025年第5期43卷 2213-2225页

作者： Su, Ye Chen, Zhuang Xu, Fang Ye, Yichen Jiang, Xiao Liu, Weichen Xie, Yiyuan Southwest Univ Coll Elect & Informat Engn Chongqing 400715 Peoples R China Nanyang Technol Univ Sch Comp Sci & Engn Singapore 639798 Singapore Southwest Univ Chongqing Key Lab Nonlinear Circuits & intelligent Chongqing 400715 Peoples R China Key Lab Networks & Cloud Comp Secur Univ Chongqing 400715 Peoples R China

Motivated by the increasing capability of artificial intelligence (AI) in solving a large class of problems, integrated photonic neural networks (PNNs) with Mach-Zehnder Interferometers (MZIs), have shown some advantages, such as low power, low latency, and high bandwidth alternatives to digitally electric neural networks. However, as the complexity of the problem being tackled grows, PNNs are accompanied by massive model sizes, necessitating significant computational and tuning power consumption. To enable the deployment of large-scale PNNs in power-constrained environments and maintain inference performance, in this paper, we target at pruning the redundant phase weights in PNNs. More specifically, we first investigate the feasibility of pruning based on phase size and point out sparse pruning schemes. Additionally, a multi-objective PNNs pruning method that trade-offs the accuracy and the tuning power consumption of networks is proposed and we solve this model with swarm optimization, named PP-MOPSO. Experimental results demonstrate that PP-MOPSO achieves a 97.3% reduction in on-chip power consumption while maintaining 90.51% inference accuracy on PNNs using the MNIST dataset. In the other case with CIFAR-10 dataset, the method achieves 94.98% power savings with a 6.39% accuracy loss.

关键词： Transmission line matrix methods Photonics Phase shifters Tuning Power demand Matrix decomposition System-on-chip Accuracy Optical network units Optical devices Low power consumption Mach-Zehnder interferometer (MZI) multi objective optimization phases pruning photonic neural networks

来源：评论

学校读者我要写书评

暂无评论

Exponential stability of delayed neural networks with delayed sampled-data inputs: An extended bilateral looped functional approach

引用

JOURNAL OF THE FRANKLIN INSTITUTE 2023年第13期360卷 9974-9995页

作者： Chang, Fei Li, Chuandong Deng, Hao Wu, Hongjuan Huan, Mingchen Southwest Univ Coll Elect & Informat Engn Chongqing Key Lab Nonlinear Circuits & Intelligent Chongqing 400715 Peoples R China Chongqing Three Gorges Univ Sch Comp Sci & Engn Chongqing 404120 Peoples R China

In this paper, the exponential stability of delayed neural networks (DNNs) with delayed sampleddata inputs is investigated via extended bilateral looped functional approach. Firstly, a new extended bilateral looped functional is constructed, which is differentiable at sampling intervals and can relax the constraints on positive definiteness when compared to traditional functionals. Then, less conservative criteria for exponential stability of DNNs with delayed sampled-data inputs expressed through linear matrix inequalities (LMIs) are obtained. Furthermore, the results are extended to T-S fuzzy DNNs with delayed sampled-data inputs, where corresponding stability conditions are likewise derived. Finally, two simulation examples are given to illustrate the validity of the main results.& COPY;2023 The Franklin Institute. Published by Elsevier Inc. All rights reserved.

关键词： Linear matrix inequalities

来源：评论

学校读者我要写书评

暂无评论

A Sparse Recovery Algorithm Using the Neurodynamic System With Predefined Time Convergence

引用

IEEE TRANSACTIONS ON circuits AND SYSTEMS II-EXPRESS BRIEFS 2023年第8期70卷 3029-3033页

作者： Zhu, Yiyue He, Xing Southwest Univ Westa Coll Chongqing 400715 Peoples R China Southwest Univ Coll Elect & Informat Engn Chongqing Key Lab Nonlinear Circuits & Intelligent Chongqing 400715 Peoples R China

Compressed sensing aims to compress sparse signals losslessly into significantly smaller samples and recover them if needed. One of the most critical approaches to recovering the sparse signal is minimizing $l_nonlinear$ -norm, the convex relaxation of $l_nonlinear$ -norm with the limit of linear measurement. However, sometimes such as the recovery of piecewise constant sparse signal, the need for controlling the convergence time has become an inevitable problem to solve. To solve this problem, the predefined time pseudo-inverse locally competitive algorithm (PPLCA) based on a nonlinear neurodynamic system is proposed in this brief, which is exceptionally effective if the signal is piecewise constant. Further, under the restricted isometry property (RIP), the predefined-time convergence rate of the proposed PPLCA is proven using Lyapunov's stability theory. Finally, the numerical simulations show that the PPLCA could converge to the stationary point within the predefined time.

关键词： Index Terms-Compressed sensing l1-norm predefined local competitive algorithm predefined time local competitive algo-rithm sliding mode control technique

来源：评论

学校读者我要写书评

暂无评论

A Fixed-Time proximal Gradient Neurodynamic Network With Time-Varying Coefficients for Composite Optimization problems and Sparse Optimization problems With Log-Sum Function

引用

IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS 2025年第5期36卷 8414-8428页

作者： Xu, Jing Li, Chuandong He, Xing Wen, Hongsong Ju, Xingxing Southwest Univ Coll Elect & Informat Engn Chongqing Key Lab Nonlinear Circuits & Intelligen Chongqing 400715 Peoples R China Xinjiang Agr Univ Sch Comp & Informat Engn Urumqi 830052 Peoples R China Sichuan Univ Coll Elect & Informat Engn Chengdu 610065 Peoples R China

This article presents a novel proximal gradient neurodynamic network (PGNN) for solving composite optimization problems (COPs). The proposed PGNN with time-varying coefficients can be flexibly chosen to accelerate the network convergence. Based on PGNN and sliding mode control technique, the proposed time-varying fixed-time proximal gradient neurodynamic network (TVFxPGNN) has fixed-time stability and a settling time independent of the initial value. It is further shown that fixed-time convergence can be achieved by relaxing the strict convexity condition via the Polyak-Lojasiewicz condition. In addition, the proposed TVFxPGNN is being applied to solve the sparse optimization problems with the log-sum function. Furthermore, the field-programmable gate array (FPGA) circuit framework for time-varying fixed-time PGNN is implemented, and the practicality of the proposed FPGA circuit is verified through an example simulation in Vivado 2019.1. Simulation and signal recovery experimental results demonstrate the effectiveness and superiority of the proposed PGNN.

关键词： Composite optimization problems (COPs) fixed-time convergence proximal gradient neurodynamic network (PGNN) signal recovery time-varying coefficients

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：