- 标题
- 摘要
- 关键词
- 实验方案
- 产品
过滤筛选
- 2018
- Conditional Random Fields (CRF)
- Convolutional Neural Network (CNN)
- Fine Classification
- Airborne hyperspectral
- Optoelectronic Information Science and Engineering
- Wuhan University
- Central South University
- Hubei University
-
[IEEE 2018 International Conference on 3D Vision (3DV) - Verona (2018.9.5-2018.9.8)] 2018 International Conference on 3D Vision (3DV) - Cross-Domain Image-Based 3D Shape Retrieval by View Sequence Learning
摘要： We propose a cross-domain image-based 3D shape retrieval method, which learns a joint embedding space for natural images and 3D shapes in an end-to-end manner. The similarities between images and 3D shapes can be computed as the distances in this embedding space. To better encode a 3D shape, we propose a new feature aggregation method, Cross-View Convolution (CVC), which models a 3D shape as a sequence of rendered views. For bridging the gaps between images and 3D shapes, we propose a Cross-Domain Triplet Neural Network (CDTNN) that incorporates an adaptation layer to match the features from different domains better and can be trained end-to-end. In addition, we speed up the triplet training process by presenting a new fast cross-domain triplet neural network architecture. We evaluate our method on a new image to 3D shape dataset for category-level retrieval and ObjectNet3D for instance-level retrieval. Experimental results demonstrate that our method outperforms the state-of-the-art approaches in terms of retrieval performance. We also provide in-depth analysis of various design choices to further reduce the memory storage and computational cost.
关键词： joint embedding space,image-based 3D shape retrieval,Cross-Domain Triplet Neural Network,Cross-View Convolution,cross-domain
更新于2025-09-23 15:21:01
-
[IEEE 2018 IEEE 31st Canadian Conference on Electrical & Computer Engineering (CCECE) - Quebec City, QC (2018.5.13-2018.5.16)] 2018 IEEE Canadian Conference on Electrical & Computer Engineering (CCECE) - Solar Forecasting Using Remote Solar Monitoring Stations and Artificial Neural Networks
摘要： The need to accurately forecast available solar irradiance is a significant issue for the power industry and poses special challenges for utilities who serve customers in isolated regions where weather forecast data may not be abundant. This paper proposes a method to forecast two hour ahead solar irradiance levels at a site in Northwestern Alberta, Canada using real-time solar irradiance measured both locally and at remote monitoring stations. This paper makes use of an Artificial Neural Network (ANN) to forecast the solar irradiance levels and uses the genetic algorithm to determine the optimal array size and positioning of solar monitoring stations to obtain the most accurate forecast from the ANN. The findings of this paper are that it is possible to use as few as five remote monitoring stations to obtain a near-peak forecasting accuracy from the algorithm and that providing adequate geospatial separation of the remote monitoring sites around the target site is more desirable than clustering the sites in the strictly upwind directions.
关键词： GHI,remote sensing,solar,PV,isolated generation,forecasting,irradiance,artificial neural network
更新于2025-09-23 15:21:01
-
[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) - Visual Tree Convolutional Neural Network in Image Classification
摘要： In image classification, Convolutional Neural Network (CNN) models have achieved high performance with the rapid development in deep learning. However, some categories in the image datasets are more difficult to distinguished than others. Improving the classification accuracy on these confused categories is benefit to the overall performance. In this paper, we build a Confusion Visual Tree (CVT) based on the confused semantic level information to identify the confused categories. With the information provided by the CVT, we can lead the CNN training procedure to pay more attention on these confused categories. Therefore, we propose Visual Tree Convolutional Neural Networks (VT-CNN) based on the original deep CNN embedded with our CVT. We evaluate our VT-CNN model on the benchmark datasets CIFAR-10 and CIFAR-100. In our experiments, we build up 3 different VT-CNN models and they obtain improvement over their based CNN models by 1.36%, 0.89% and 0.64%, respectively.
关键词： Image Classification,Convolutional Neural Network,Confusion Visual Tree,Deep Learning
更新于2025-09-23 15:21:01
-
[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) - Semi-supervised convolutional neural networks with label propagation for image classification
摘要： Over the past several years, deep learning has achieved promising performance in many visual tasks, e.g., face verification and object classification. However, a limited number of labeled training samples existing in practical applications is still a huge bottleneck for achieving a satisfactory performance. In this paper, we integrate class estimation of unlabeled training data with deep learning model which generates a novel semi-supervised convolutional neural network (SSCNN) trained by both the labeled training data and unlabeled data. In the framework of SSCNN, the deep convolution feature extraction and the class estimation of the unlabeled data are jointly learned. Specifically, deep convolution features are learned from the labeled training data and unlabeled data with confident class estimation. After the deep features are obtained, the label propagation algorithm is utilized to estimate the identities of unlabeled training samples. The alternative optimization of SSCNN makes the class estimation of unlabeled data more and more accurate due to the learned CNN feature more and more discriminative. We compared the proposed SSCNN with some representative semi-supervised learning approaches on MINIST and Cifar-10 databases. Extensive experiments on landmark databases show the effectiveness of our semi-supervised deep learning framework.
关键词： convolutional neural network,semi-supervised learning,label propagation
更新于2025-09-23 15:21:01
-
A Comparative Study on Neural Network Based Controllers Used in Grid-Interactive Solar System
摘要： This paper proposes a comparative assessment of NN based current controllers followed by hardware validation towards power quality improvement in grid-interactive VSI controlled solar system. The steady state errors, transient disturbances and high current harmonic effects encountered in conventional linear PI and PR controllers are nullified by employing intelligent adaptive current controller. Three adaptive current control strategies viz. ADALINE-LMS, ALLMS, and VLAS-LMS are identified by using artificial neural network topology having been controlled by different weight-regulating algorithms which helps in minimizing current harmonics generated by the widespread use of VSI, non-linear loads, faults and uncertain polluted grid. A comprehensive comparative study is driven from the proposed adaptive controller’s stability and convergence criterion, current magnitudes calculated at different power zones, % overshoot, settling time and power quality and analyzed under numerous operating conditions. From the comparative assessment performed in between conventional and intelligent current controllers, it is confirmed that the intelligent control technique performs best under the non-linear loads and transient conditions whereas all the controllers perform equally good under the linear loads. Proposed methods are simulated in MATLAB/Simulink and their effectiveness is compared in terms of time responses, stability and low-order current harmonics compensation capability. The robustness of the intelligent current controller is established through the experimental performances using dSPACE RTI 1202.
关键词： grid-interactive solar,power quality,stability,dSPACE,neural network (NN),current controllers (cc),time response
更新于2025-09-23 15:21:01
-
Automatic lung segmentation in low-dose chest CT scans using convolutional deep and wide network (CDWN)
摘要： Computed tomography (CT) imaging is the preferred imaging modality for diagnosing lung-related complaints. Automatic lung segmentation is the most common prerequisite to develop a computerized diagnosis system for analyzing chest CT images. In this paper, a convolutional deep and wide network (CDWN) is proposed to segment lung region from the chest CT scan for further medical diagnosis. Earlier lung segmentation techniques depend on handcrafted features, and their performance relies on the features considered for segmentation. The proposed model automatically segments the lung from complete CT scan in two laps: (1) learning the required ?lters to extract hierarchical feature representations at convolutional layers, (2) dense prediction with spatial features through learnable deconvolutional layers. The model has been trained and evaluated with low-dose chest CT scan images on LIDC-IDRI database. The proposed CDWN reaches the average Dice coef?cient of 0.95 and accuracy of 98% in segmenting the lung regions from 20 test images and maintains consistent results for all test images. The experimental results con?rm that the proposed approach achieves a superior performance compared to other state-of-the-art methods for lung segmentation.
关键词： Medical imaging,Image processing and analysis,Deep learning,Automatic lung segmentation,Convolutional neural network
更新于2025-09-23 15:21:01
-
Functional-Link Neural Network for Nonlinear Equalizer in Coherent Optical Fiber Communications
摘要： We propose and experimentally demonstrate a simple nonlinear equalizer based on functional-link neural network (FLNN). The nonlinear stochastic mapping enables FLNN to serve as a nonlinear network, so we construct an FLNN with the signals from the two polarizations and the mapped features as input to combat the fiber nonlinearity in coherent optical transmission systems. The FLNN can use the Moore-Penrose generalized inverse or the ridge regression to solve the weights, which can speed up the training process, and avoid the iterative and time-consuming training process that exist universally in most of the deep neural networks. We also extend the FLNN to the multi-channel transmissions. All of the received signals from different channels are stretched as the input and then we use a joint FLNN to extract features and equalize the nonlinear distortions. We conduct simulations and experiments to verify the proposed scheme. In the simulation and experiment, we transmit a 128 Gb/s polarization division multiplexed 16-QAM (PDM-16-QAM) signal over 1000-km and 600-km standard single mode fiber (SSMF), respectively. Both the simulation and experimental results show that the FLNN has similar performance as deep neural network (DNN), which can improve the transmission performance in the nonlinear region. Moreover, the FLNN can avoid the gradient dissipation and local minimum problems in DNN, which simplify the training process. We also extend the proposed scheme in a five-channel (5 × 160 Gb/s) multiplexed transmission system. In simulation, we use joint FLNN and joint DNN to compensate the nonlinear distortions, respectively. We find that the BERs of the five channels can be below 7% HD-FEC with nonlinear equalizer.
关键词： enhancement node,functional-link neural network,Optical communication,deep neural network,fiber nonlinearity
更新于2025-09-23 15:19:57
-
Embedding physics domain knowledge into a Bayesian network enables layer-by-layer process innovation for photovoltaics
摘要： Process optimization of photovoltaic devices is a time-intensive, trial-and-error endeavor, which lacks full transparency of the underlying physics and relies on user-imposed constraints that may or may not lead to a global optimum. Herein, we demonstrate that embedding physics domain knowledge into a Bayesian network enables an optimization approach for gallium arsenide (GaAs) solar cells that identifies the root cause(s) of underperformance with layer-by-layer resolution and reveals alternative optimal process windows beyond traditional black-box optimization. Our Bayesian network approach links a key GaAs process variable (growth temperature) to material descriptors (bulk and interface properties, e.g., bulk lifetime, doping, and surface recombination) and device performance parameters (e.g., cell efficiency). For this purpose, we combine a Bayesian inference framework with a neural network surrogate device-physics model that is 100× faster than numerical solvers. With the trained surrogate model and only a small number of experimental samples, our approach reduces significantly the time-consuming intervention and characterization required by the experimentalist. As a demonstration of our method, in only five metal organic chemical vapor depositions, we identify a superior growth temperature profile for the window, bulk, and back surface field layer of a GaAs solar cell, without any secondary measurements, and demonstrate a 6.5% relative AM1.5G efficiency improvement above traditional grid search methods.
关键词： Bayesian network,GaAs solar cells,photovoltaics,neural network surrogate model,process optimization
更新于2025-09-23 15:19:57
-
Deep learning based automatic defect identification of photovoltaic module using electroluminescence images
摘要： The maintenance of large-scale photovoltaic (PV) power plants is considered as an outstanding challenge for years. This paper presented a deep learning-based defect detection of PV modules using electroluminescence images through addressing two technical challenges: (1) providing a large number of high-quality Electroluminescence (EL) image generation method for the limit of EL image samples; and (2) an efficient model for automatic defect classification with the generated EL image. The EL image generation approach combines traditional image processing technology and GAN characteristics. It can produce a large number of EL image samples with high resolution using a limited number of samples. Then, a convolution neural network (CNN) based model for the automatic classification of defects in an EL image is presented. CNN is used to extract the deep feature of the EL image. It can greatly increase the accuracy and efficiency of PV modules inspection and health management in comparison with the other solutions. The proposed solution is assessed through extensive experiments by using the existing machine learning models, VGG16, ResNet50, Inception V3 and MobileNet, as the comparison benchmarks. The numerical results confirm that the proposed deep learning-based solution can carry out efficient and accurate defect detection automatically using the electroluminescence images.
关键词： Automatic defect classification,Electroluminescence Images,Generative adversarial network,Convolution neural network
更新于2025-09-23 15:19:57
-
[IEEE 2019 8th International Conference on Renewable Energy Research and Applications (ICRERA) - Brasov, Romania (2019.11.3-2019.11.6)] 2019 8th International Conference on Renewable Energy Research and Applications (ICRERA) - A Study of Introduction of the Photovoltaic Generation System to Conventional Railway
摘要： This paper presents a voice conversion (VC) method that utilizes the recently proposed probabilistic models called recurrent temporal restricted Boltzmann machines (RTRBMs). One RTRBM is used for each speaker, with the goal of capturing high-order temporal dependencies in an acoustic sequence. Our algorithm starts from the separate training of one RTRBM for a source speaker and another for a target speaker using speaker-dependent training data. Because each RTRBM attempts to discover abstractions to maximally express the training data at each time step, as well as the temporal dependencies in the training data, we expect that the models represent the linguistic-related latent features in high-order spaces. In our approach, we convert (match) features of emphasis for the source speaker to those of the target speaker using a neural network (NN), so that the entire network (consisting of the two RTRBMs and the NN) acts as a deep recurrent NN and can be fine-tuned. Using VC experiments, we confirm the high performance of our method, especially in terms of objective criteria, relative to conventional VC methods such as approaches based on Gaussian mixture models and on NNs.
关键词： recurrent temporal restricted Boltzmann machine (RTRBM),voice conversion,speaker specific features,recurrent neural network,Deep Learning
更新于2025-09-23 15:19:57