- 标题
- 摘要
- 关键词
- 实验方案
- 产品
-
[IEEE 2018 IEEE International Conference on Computational Electromagnetics (ICCEM) - Chengdu, China (2018.3.26-2018.3.28)] 2018 IEEE International Conference on Computational Electromagnetics (ICCEM) - Coupling Matrix Extraction for Microwave Filter Design Using Neural Networks
摘要： In this paper, a novel coupling matrix extraction method for microwave ?lters is proposed. The neural networks (NNs) are introduced to extract the coupling matrix from the simulated responses. Compared to the traditional Cauchy method or vector ?tting with complicated derivations, the new method inherits the high speed of the trained NNs and can extract the coupling matrix of target S-parameters with high accuracy simultaneously. Finally, the coupling matrix of a fourth-order bandpass ?lter is extracted by using NNs. The numerical results validate the effectiveness of the proposed method.
关键词： Error back propagation neural networks (BPNNs),parameters extraction,microwave ?lters,coupling matrix
更新于2025-09-10 09:29:36
-
[IEEE 2018 20th International Conference on Transparent Optical Networks (ICTON) - Bucharest (2018.7.1-2018.7.5)] 2018 20th International Conference on Transparent Optical Networks (ICTON) - Impact of Selected Input Features for Lightpath Feasibility Validation Using Artificial Neural Networks
摘要： The new advents of 5G and Internet of Things (IoT) will impact the traf?c, both in volume and dynamicity, at unprecedented rates. As a result, optical transport networks should become more responsive to the traf?c changes as well as to operate more closely to optimality. Therefore, the implementation of a self-driving network is being proposed as a way to achieve these targets. One of the key challenges in this environment is the automatic provisioning of lightpaths. In order to provision a lightpath, Quality of Transmission (QoT) needs to be estimated, which involves complex and time consuming computations. This work proposes the use of arti?cial neural networks (ANN) to speed up lightpath feasibility validation without performing full validation per request (slow) nor keeping a full database of feasible lightpaths (memory consuming). Moreover, we evaluate the impact of input features selection and number of neurons in the obtained accuracy.
关键词： transport networks,arti?cial neural networks,machine learning,Quality of Transmission
更新于2025-09-09 09:28:46
-
[IEEE IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium - Valencia, Spain (2018.7.22-2018.7.27)] IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium - A Deep Network Approach to Multitemporal Cloud Detection
摘要： We present a deep learning model with temporal memory to detect clouds in image time series acquired by the Seviri imager mounted on the Meteosat Second Generation (MSG) satellite. The model provides pixel-level cloud maps with related confidence and propagates information in time via a recurrent neural network structure. With a single model, we are able to outline clouds along all year and during day and night with high accuracy.
关键词： convolutional neural networks,Seviri,deep learning,Cloud detection,recurrent neural networks
更新于2025-09-09 09:28:46
-
[IEEE 2018 24th International Conference on Pattern Recognition (ICPR) - Beijing, China (2018.8.20-2018.8.24)] 2018 24th International Conference on Pattern Recognition (ICPR) - Reservoir Computing with Untrained Convolutional Neural Networks for Image Recognition
摘要： Reservoir computing has attracted much attention for its easy training process as well as its ability to deal with temporal data. A reservoir computing system consists of a reservoir part represented as a sparsely connected recurrent neural network and a readout part represented as a simple regression model. In machine learning tasks, the reservoir part is fixed and only the readout part is trained. Although reservoir computing has been mainly applied to time series prediction and recognition, it can be applied to image recognition as well by considering an image data as a sequence of pixel values. However, to achieve a high performance in image recognition with raw image data, a large-scale reservoir including a large number of neurons is required. This is a bottleneck in terms of computer memory and computational cost. To overcome this bottleneck, we propose a new method which combines reservoir computing with untrained convolutional neural networks. We use an untrained convolutional neural network to transform raw image data into a set of smaller feature maps in a preprocessing step of the reservoir computing. We demonstrate that our method achieves a high classification accuracy in an image recognition task with a much smaller number of trainable parameters compared with a previous study.
关键词： Reservoir computing,Image recognition,Untrained networks,Convolutional neural networks
更新于2025-09-09 09:28:46
-
DuFiNet: Architectural Considerations and Physical Layer Studies of an Agile and Cost-Effective Metropolitan Area Network
摘要： Metro networks are becoming considerably important to support existing and future 5G services and they are facing new challenges. To address the new requirements, we present a novel, highly flexible and low-cost, filterless Metro network that overcomes existing architectural limitations of filterless solutions. In particular, we detail architectural aspects of this architecture at network and node level as well as the protection strategy. Finally we introduce a novel physical layer modeling tailored for Metro, to assess the capacity and scalability of the transport platform.
关键词： physical layer modeling,filterless network architecture,metropolitan area networks,optical communications
更新于2025-09-09 09:28:46
-
Marginless Operation of Optical Networks
摘要： Considering flexible technologies available nowadays, operating optical networks much closer to their physical capacities is very tempting but necessarily requires efficient network automation. To achieve this, the two main challenges are handling failures, and accurately predicting performance in dynamic environments. We experimentally demonstrate the ability of the ORCHESTRA solution for early detection and localization of failures, to preventively mitigate their impact and thus guarantee smooth network operation. Then, leveraging machine learning for live performance estimation and closed-loop software-defined network (SDN) control, we demonstrate a fully automated reconfiguration of marginless connections undergoing critical performance variations over 228km of field-deployed fiber.
关键词： Optical networks,cross-layer optimization,failure localization,marginless operations,soft-failures
更新于2025-09-09 09:28:46
-
[IEEE 2018 International Conference on 3D Vision (3DV) - Verona (2018.9.5-2018.9.8)] 2018 International Conference on 3D Vision (3DV) - SeeThrough: Finding Objects in Heavily Occluded Indoor Scene Images
摘要： Discovering 3D arrangements of objects from single indoor images is important given its many applications such as interior design and content creation for virtual environments. Although heavily researched in the recent years, existing approaches break down under medium to heavy occlusion as the core image-space region detection module fails in absence of directly visible cues. Instead, we take into account holistic contextual 3D information, exploiting the fact that objects in indoor scenes co-occur mostly in typical configurations. First, we use a neural network trained on real indoor annotated images to extract 2D keypoints, and feed them to a 3D candidate object generation stage. Then, we solve a global selection problem among these candidates using pairwise co-occurrence statistics discovered from a large 3D scene database. We iterate the process allowing for candidates with low keypoint response to be incrementally detected based on the location of the already discovered nearby objects. We demonstrate significant performance improvement over combinations of state-of-the-art methods, especially for scenes with moderately to severely occluded objects. Code and data available at http://geometry.cs.ucl.ac.uk/projects/2018/seethrough.
关键词： 3D vision,indoor scenes,object detection,occlusion,neural networks
更新于2025-09-09 09:28:46
-
[IEEE 2018 IEEE International Conference on Smart Energy Grid Engineering (SEGE) - Oshawa, ON (2018.8.12-2018.8.15)] 2018 IEEE International Conference on Smart Energy Grid Engineering (SEGE) - Effects of Centralized Battery Storage Placement in Low-Voltage Residential Distribution Networks with High Photovoltaic Penetration
摘要： In this paper, power flow simulation is used to investigate how centralized battery storage can aide in mitigating under-voltage conditions in highly PV penetrated residential network. It also investigates how the placement of the battery storage relates to the rating of the storage used. It is found the reduction in storage rating is quadratically proportional to the distance along the low-voltage distribution feeder.
关键词： photovoltaic generation,voltage regulation,battery storage,low-voltage networks
更新于2025-09-09 09:28:46
-
[IEEE 2018 IEEE 16th International Conference on Software Engineering Research, Management and Applications (SERA) - Kunming (2018.6.13-2018.6.15)] 2018 IEEE 16th International Conference on Software Engineering Research, Management and Applications (SERA) - Motion Deblurring via Using Generative Adversarial Networks for Space-Based Imaging
摘要： In some missions of NanoSats, we find images captured are disturbed by motion blur which caused under the situation that NanoSats work in low-earth orbit at high speeds. In this paper, we address the problem of deblurring images degraded due to space-based imaging system shaking or movements of observing targets. We propose a motion deblurring strategy via using Generative Adversarial Networks(GAN) to realize an end-to-end image processing without kernel estimation in orbit. We combine Wasserstein GAN(WGAN) and loss function based on adversarial loss and perceptual loss to optimize the result of deblurred image. The experimental results on the two different datasets prove the feasibility and effectiveness of the proposed strategy which outperforms the state-of-the-art blind deblurring algorithms using for remote sensing images both quantitatively and qualitatively.
关键词： Space-Based Imaging,Generative Adversarial Networks,NanoSats,Motion Deblurring
更新于2025-09-09 09:28:46
-
[IEEE 2018 24th International Conference on Pattern Recognition (ICPR) - Beijing, China (2018.8.20-2018.8.24)] 2018 24th International Conference on Pattern Recognition (ICPR) - Improving Image Classification Performance with Automatically Hierarchical Label Clustering
摘要： Image classi?cation is a common and foundational problem in computer vision. In traditional image classi?cation, a category is assigned with single label, which is dif?cult for networks to learn better features. On the contrary, hierarchical labels can depict the structure of categories better, which helps network to learn more hierarchical features and improve the classi?cation performance. Though many datasets contain images with multi-labels, the labels in these datasets usually lack of hierarchy. To overcome this problem, we propose a new method to improve image classi?cation performance with Automatically Hierarchical Label Clustering (AHLC). Firstly, AHLC calculates the similarity between each pair of original categories by how easily they are misclassi?ed with a pre-trained classi?er. Secondly, AHLC obtains hierarchical labels by merging similar categories using hierarchical clustering. Finally, AHLC trains a new classi- ?er with hierarchial labels to improve the original classi?cation performance. We evaluate our method on MNIST and CIFAR-100 datasets and the results demonstrate the superiority of our method. The main contribution of this work is that we can simply improve an existing classi?cation network by AHLC without extra information or heavy architecture redesign.
关键词： hierarchical labels,AHLC,Image classification,deep learning,convolutional neural networks
更新于2025-09-09 09:28:46