检索结果-内蒙古大学图书馆

International Conference on Image, Vision and Computing (ICIVC)

作者： Jie Ding Jianxin Chen Anqi Zhou Zhimin Chen Key Lab of Broadband Wireless Communication and Sensor Network Technology Nanjing University of Posts and Telecommunications Ministry of Education Nanjing China

ISBN: (纸本)9781728123264

With the appearance of RGB-D cameras, 3D reconstruction becomes convenient. Previous studies focused on the objects with strong structural features. But for the objects with weak structure, such as rotationally symmetric objects. There is no solution. In this paper, we propose a method to model the rotationally symmetric objects. In this method, we scan the object in one view and the point cloud is projected to the 2D space. Then, we extract the boundary of the point cloud. And the Least Square is adopted to fit the extracted boundary to improve the accuracy of the boundary. At last, we rotate the fitting boundary and get the model. To verify the effectiveness of our method, several experiments have been performed to evaluate the error of the models. In addition, we use it to reconstruct the object with the hole on one side. And the results show the efficiency and computational complexity of our method.

关键词： Three-dimensional displays Fitting Computational modeling Solid modeling Mathematical model Strain Cameras

来源：评论

学校读者我要写书评

暂无评论

QoE Optimization for Traffic Offloading from LTE to WiFi

QoE Optimization for Traffic Offloading from LTE to WiFi

引用

IEEE Global Conference on Consumer Electronics (GCCE)

作者： Feng Tian Yue Yu Dapeng Li Jingwu Cui Yue Dong Key Lab of Broadband Wireless Communication and Sensor Network Technology (Nanjing University of Posts and Telecommunications) Ministry of Education Nanjing China

ISBN: (数字)9781728135755

ISBN: (纸本)9781728135762

In this paper, a novel QoE (Quality of Experience) optimization mechanism is proposed for the coexistence between LTE and WiFi networks, where traffic offloading is enabled from LTE to WiFi. Firstly, the throughput models of LTE and WiFi are analyzed, respectively. And then the users' satisfaction function is introduced to measure QoE. An optimal problem that maximizing the QoE of LTE network with guaranteeing the QoE of WiFi network, is formulated and solved by transformation and simplification. Simulation results indicate that traffic offloading can greatly improve the QoE of LTE network.

关键词： wireless fidelity Long Term Evolution Quality of experience Optimization Throughput wireless communication Interference

来源：评论

学校读者我要写书评

暂无评论

Moving Object Distance Estimation Method Based on Target Extraction with a Stereo Camera

Moving Object Distance Estimation Method Based on Target Ext...

引用

International Conference on Image, Vision and Computing (ICIVC)

作者： Jingyi Zhang Jianxin Chen Qingyu Lin Liming Cheng Key Lab of Broadband Wireless Communication and Sensor Network Technology Nanjing University of Posts and Telecommunications Ministry of Education Nanjing China

ISBN: (纸本)9781728123264

Real-time distance estimation with high accuracy is crucial to robot applications such as navigation and obstacle negotiation. However, considering the truth that high accuracy brings huge computation, the previous work either accelerates the distance estimation process through hardware, or only estimates the static objects, or sacrifices accuracy to realize real-time. In order to improve the trade-off between speed and accuracy, in this paper we propose a stereovision-based method that integrates the object detection and tracking into stereo matching. We utilize kernelized correlation filter (KCF) to track the moving object, and then extract target regions from image pair as input images to compute disparity map. In addition, we update each pixel's disparity by a simple but effective approach to further refine disparity map. Eventually, distance is calculated with disparity by triangulation. Experimental results show that our proposed method is suitable for dynamic environment, the extraction method improves the speed of the entire system, and the average accuracy of distance estimation can achieve over 95% from 88 cm to 300 cm when baseline is 151 mm.

关键词： Distance estimation Stereo vision KCF tracking Target region extraction Disparity refinement Triangulation Stereo cameras triangulation Stereoviews extraction of object Object detection object distance Static objects

来源：评论

学校读者我要写书评

暂无评论

Gradient-Free optimization algorithm for single-qubit quantum classifier

arXiv

引用

arXiv 2022年

作者： Zhang, Anqi He, Xiaoyun Zhao, Shengmei Nanjing210003 China Key Lab of Broadband Wireless Communication and Sensor Network Technology Ministry of Education Nanjing210003 China

In the paper, a gradient-free optimization algorithm for single-qubit quantum classifier is proposed to overcome the effects of barren plateau caused by quantum devices. A rotation gate RX() is applied on a single-qubit binary quantum classifier, and the training data and parameters are loaded into with the form of vector-multiplication. The cost function is decreased by finding the value of each parameter that yield the minimum expectation value of measuring the quantum circuit. The algorithm is performed iteratively for all parameters one by one, until the cost function satisfies the stop condition. The proposed algorithm is demonstrated for a classification task and is compared with that using Adam optimizer. Furthermore, the performance of the single-qubit quantum classifier with the proposed gradient-free optimization algorithm is discussed when the rotation gate in quantum device is under different noise. The simulation results show that the single-qubit quantum classifier with proposed gradient-free optimization algorithm can reach a high accuracy faster than that using Adam optimizer. Moreover, the proposed gradient-free optimization algorithm can quickly completes the training process of the single-qubit classifier. Additionally, the single-qubit quantum classifier with proposed gradient-free optimization algorithm has a good performance in noisy environments. Copyright © 2022, The Authors. All rights reserved.

关键词： Cost functions

来源：评论

学校读者我要写书评

暂无评论

Ternary Weighted networks with Equal Quantization Levels

Ternary Weighted Networks with Equal Quantization Levels

引用

Asia-Pacific Conference on communications

作者： Yuanyuan Chang Xiaofu Wu Suofei Zhang Jun Yan Key Lab of Ministry of Education in Broadband Wireless Communication and Sensor Network Technology Nanjing University of Posts and Telecommunications Nanjing China

ISBN: (数字)9781728136790

ISBN: (纸本)9781728136806

Recent progress in deep convolutional neural networks has considerably changed the landscape of speech recognition, computer vision, natural language processing and so on. However, limited by the large number of parameters, the memory space and high computational complexity, it is a challenging task to deploy the deep neural network model in embedded system. To solve this problem, we propose Equal Trained Ternary Quantization (ETTQ), a ternary quantization method by improving Trained Ternary Quantization, which uses only a full-precision scaling coefficient for each layer, and quantize the weights to three levels. These positive and negative weights have same absolute values that are trainable parameters. Experiments show that the performance of ETTQ is only slightly worse than TTQ, but can converge faster and more stable during the training over CIFAR-10 and CIFAR-100 datasets.

关键词： Quantization (signal) Neural networks Training Computational modeling Solid modeling Computer vision Backpropagation

来源：评论

学校读者我要写书评

暂无评论

Geometry Based LM of Robot to Imitate Human Motion with Kinect

Geometry Based LM of Robot to Imitate Human Motion with Kine...

引用

International Conference on Image, Vision and Computing (ICIVC)

作者： Wei Guo Jianxin Chen Ming Zhang Zhaolai Pan Ministry of Education Key Lab of Broadband Wireless Communication and Sensor Network Technology (Nanjing University of Posts and Telecommunications) Nanjing China

ISBN: (纸本)9781728123264

Humanoid robots are easily accepted by people because they have a body shape and appearance close to human. Humanoid robots have many application scenarios and great application potentials. Nevertheless, it is not easy to realize robot simulation due to the different physical structures of human bodies and robots. In this paper, we propose a geometry based LM algorithm for the robot to imitate the human motion of upper limb in real time. This method greatly improves the joint angle similarity under the premise of ensuring the similarity of the end trajectories. We designed a comprehensive and unrestricted motion simulation experiment to verify this method.

关键词： Kinect Inverse kinematic Nao Levenberg-Marquardt Geometry game player geometry Robot Human motion Humanoid robots Human Body

来源：评论

学校读者我要写书评

暂无评论

Variational Quantum Domain Adaptation

arXiv

引用

arXiv 2023年

作者： Wu, Chunhui Pei, Junhao Wu, Yihua Zhao, Shengmei Nanjing210003 China Key Lab of Broadband Wireless Communication and Sensor Network Technology Ministry of Education Nanjing210003 China

Quantum machine learning is an important application of quantum computing in the era of noisy intermediate-scale quantum devices. Domain adaptation is an effective method for addressing the distribution discrepancy problem between the training data and the real data when the neural network model is deployed. In this paper, a variational quantum domain adaptation method is proposed by using a quantum convolutional neural network, together with a gradient reversal module, and two quantum fully connected layers, named variational quantum domain adaptation (VQDA). The simulations on the local computer and IBM Quantum Experience (IBM Q) platform by Qiskit show the effectiveness of the proposed method. The results demonstrate that, compared to its classical corresponding domain adaptation method, VQDA achieves an average improvement of 4 % on the accuracy for MNIST → USPS domain transfer under the same parameter scales. Similarly, for SYNDigits → SVHN domain transfer, VQDA achieves an average improvement of 2 % on the accuracy under the same parameter scales. Copyright © 2023, The Authors. All rights reserved.

关键词： Neural network models

来源：评论

学校读者我要写书评

暂无评论

Heart Rate Detection with the Off-the-Shelf Camera: Static to Non-Static

Heart Rate Detection with the Off-the-Shelf Camera: Static t...

引用

IEEE International Conference on Signal and Image Processing (ICSIP)

作者： Rui Chen Jianxin Chen LiMing Cheng XiangJun Huang Key Lab of Broadband Wireless Communication and Sensor Network Technology Nanjing University of Posts and Telecommunications Ministry of Education Nanjing China

ISBN: (数字)9781728168968

ISBN: (纸本)9781728168975

Imaging photo-plethysmography (IPPG) technique has been proven to detect heart rate (HR) with a camera. Currently it might be used in a rigid scenario, such as the static subject, the short measurement distance and good illumination. In this paper, we try to relax these constraints by extending the measurement distance and applying it to the non-static subject with IPPG technique using the off-the-shelf camera. To achieve it, the landmark detection method is used to locate the facial region of interest (ROI) for the subject during movements. On this ROI, R/G/B channel signals are extracted with the spatial average of pixels. Then we use a skin appearance model (CHROME) to remove the noise, and HR is estimated from the output of the CHROME through the power spectrum. Experimental results showed that our method works effectively in both static and non-static scenarios. In static scenario, the measurement distance can extend to 4 meters, while the accuracy can reach 86% compared with the results from a finger detection equipment. In non-static scenario, our method achieve better results than previous studies.

关键词： Heart rate Current measurement Predictive models Cameras Skin Velocity measurement Motion measurement

来源：评论

学校读者我要写书评

暂无评论

Face Recognition Based on Two-stage CNN Combined with Transfer Learning

Face Recognition Based on Two-stage CNN Combined with Transf...

引用

International Conference on Image, Vision and Computing (ICIVC)

作者： Anqi Zhou Jianxin Chen Jie Ding Zhaolai Pan Key Lab of Broadband Wireless Communication and Sensor Network Technology Nanjing University of Posts and Telecommunications Ministry of Education Nanjing China

ISBN: (纸本)9781728123264

In recent years, pattern recognition has received wide attention, especially face recognition. Research shows that traditional methods have limits in face recognition under unrestricted environment because of illumination, facial expression, pose and so on. Deep learning is chosen to solve these problems. In this paper, we propose a complete method for face recognition under unrestricted environment by deep learning. First, we design a novel two-stage convolutional neural network for face detection and alignment. The first stage of the network aims at searching face windows and their bounding box regression vector. The second stage merges facial candidates which are covered highly. Then, the network finishes the task of face detection and alignment whose outputs are face regions and the position of the five facial landmarks. It achieves great performance on angle of faces in image and obtains accuracy of 96.1% on AFW and 98.9% on FDDB. Then faces are sent to convolutional neural network for transfer learning because of limited training samples. On the testing dataset, the accuracy of face recognition reaches to 95.3%, comparing with 89.1% of the original network. What' more, our method is better than the traditional algorithms and general CNN.

关键词： Face detection Face recognition Deep learning Transfer learning Facial Recognition Transfer (Psychology) AUXILIARY FEEDWATER

来源：评论

学校读者我要写书评

暂无评论

A Reliable Hand Tracking Method Using Kinect

A Reliable Hand Tracking Method Using Kinect

引用

International Conference on Image, Vision and Computing (ICIVC)

作者： Qinyu Lin Jianxin Chen Jiangyi Zhang Ling Yao Ministry of Education Key Lab of Broadband Wireless Communication and Sensor Network Technology Nanjing University of Posts and Telecommunications Nanjing China

ISBN: (纸本)9781728123264

Hand tracking has been focused over years in the field of human-computer interaction. But there are still some challenges such as reliability and accuracy due to the highly articulated and deformable of hand. In this paper we propose a real time hand tracking method with the depth information of the Kinect sensor. In this method, we preprocessed the depth information with a pixel filter to deal with the noise. Then we isolated the hand region from the depth image by using the KCF (Kernelized Correlation Filters) to the confined minimum depth area, which can release the limitation of hand movement range. Then the forearm region is segmented with assistance of PCA (Principle Component Analysis). And, to accurately locate the fingertip position, we integrate the maximum geodesic distance to hand centroid into EIC (Extended Inscribed Circle). And by labelling each finger to avoid mistaken fingertip determination. Experimental results verify the accuracy and reliability of our solution.

关键词： Real-time systems Human computer interaction Tracking Covariance matrices Principal component analysis Wrist Reliability

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：