教学课程资源库 更多>>

关键词： capsule network   hyperspectral   image classification   deep learning   possibility density  

摘要： Deep learning techniques have boosted the performance of hyperspectral image (HSI) classification. In particular, convolutional neural networks (CNNs) have shown superior performance to that of the conventional machine learning algorithms. Recently, a novel type of neural networks called capsule networks (CapsNets) was presented to improve the most advanced CNNs. In this paper, we present a modified two-layer CapsNet with limited training samples for HSI classification, which is inspired by the comparability and simplicity of the shallower deep learning models. The presented CapsNet is trained using two real HSI datasets, i.e., the PaviaU (PU) and SalinasA datasets, representing complex and simple datasets, respectively, and which are used to investigate the robustness or representation of every model or classifier. In addition, a comparable paradigm of network architecture design has been proposed for the comparison of CNN and CapsNet. Experiments demonstrate that CapsNet shows better accuracy and convergence behavior for the complex data than the state-of-the-art CNN. For CapsNet using the PU dataset, the Kappa coefficient, overall accuracy, and average accuracy are 0.9456, 95.90%, and 96.27%, respectively, compared to the corresponding values yielded by CNN of 0.9345, 95.11%, and 95.63%. Moreover, we observed that CapsNet has much higher confidence for the predicted probabilities. Subsequently, this finding was analyzed and discussed with probability maps and uncertainty analysis. In terms of the existing literature, CapsNet provides promising results and explicit merits in comparison with CNN and two baseline classifiers, i.e., random forests (RFs) and support vector machines (SVMs).

关键词： plasma target interaction   plasma dielectric interaction   plasma jet   electric field   surface charges   Mueller polarimetry   electro-optic crystals  

摘要： Numerical simulations and experiments are performed to better understand the interaction between a pulsed helium plasma jet and a dielectric target. The focus of this work lies on the volume and surface charge influence on the electric field distribution. Experimentally, the electric field due to surface charges is measured inside an electro-optic target under exposure of a plasma jet, using the optical technique called Mueller polarimetry. For the first time, the time-resolved spatial distributions of both the axial and radial components of electric field inside the target are obtained simultaneously. A 2D fluid model is used in a complementary way to the experiments in order to study separately the contribution of volume charges and surface charges to the spatio-temporal evolutions of the electric field during the plasma-surface interaction. The experimental investigation shows that the average axial and radial components of electric field inside the dielectric target, only due to surface charges, are lower than generally reported for electric field values in the plasma plume. Thanks to the phenomenological comparison with experiments, simulations show that during the plasma-surface interaction two effects sequentially determine the electric field inside the target: firstly, a relatively high electric field is observed due to the proximity of the ionization front;afterwards, in longer timescales, lower electric fields are induced due to the contribution of both leftover volume charges close to the target and surface charges deposited on its surface. The experimental technique provides a unique way to examine this second phase of the plasma-surface interaction.

关键词： N mineralization   Nematodes   Soil biome   Microorganismos de la montana   Local effect microorganisms  

摘要： Culturing and using effective microorganisms from locally derived microorganisms has become a common practice employed by farmers throughout Central America as well as parts of South America and Southeast Asia to control foul odors and flies around compost piles and livestock facilities. These local, effective microorganisms (LEM) are applied in a concentrated solution fermented from carbohydrate-rich substrates inoculated with locally-sourced, actively decomposing leaf litter. Despite its growing prevalence among small-scale farmers in some parts of the world, there is little published research that explores the beneficial or detrimental effects of changing the biome of animal-based nutrients and the soils to which LEM are applied. The objectives of this study were to evaluate the effect of combining LEM with swine effluent to fertilize annual ryegrass (Lolium multiflontm) on plant-available nitrogen, nematode community structure, and forage productivity. Forage performance was good across all of the treatments - with relative feed quality (RFQ) values above 200 and 100 in the March and May harvests respectively. In the first, second and third sampling dates, soils from 0 to 5 cm depth showed more nitrogen mineralized in LEM-treated soils than in control soils (p = 0.178, p < 0.0001 and p = 0.027) respectively. After nutrient applications, soil-nematode community structure (0-10 cm) at the trophic group level changed over time. Most changes were similar between treatments, however, plant parasitic nematode populations were found to be significantly higher in LEM-treated plots in all but one of the sampling dates. LEM is gaining popularly as a natural amendment and although we have shown that it does increase N availability and changes trophic-level nematode community structure, more research is needed to optimize the potential of these benefits.

关键词： Object-based classifiers   Pixel-based classifiers   Landsat-8 imagery   Training sample size  

摘要： In this study, we used Landsat-8 imagery to test object- and pixel-based image classification approaches in an urban fringe area. For object-based classification, we applied four machine learning classifiers: decision tree (DT), naive Bayes (NB), random trees (RT), and support vector machine (SVM). For pixel-based classification, we utilized the maximum likelihood classifier (MLC). Specifically, we explored the influence of repeated sampling on classification results with different training sample sizes. We found that (1) except the overall accuracy of NB, those of the other four classifiers increased as the training sample size increased;(2) repeated sampling had a significant effect on classification accuracy, especially for the DT and NB classifiers;and (3) SVM achieved the best classification accuracy. In addition, the performance of the object-based classifiers was superior to that of the pixel-based classifier. The results of this study can provide guidance on the training sample size and classifier selection.

关键词： Visual aesthetic quality assessment   Aesthetic understanding   Sample-specific weighting  

摘要： Currently image aesthetic estimation using deep learning has achieved great success compared with the traditional methods by hand-crafted features. Similar to recognition problem, aesthetic estimation categorizes images into visually appealing or not. Nevertheless, it is desirable to understand why certain images are visually more appealing, in specific, which part of the image is contributing to the aesthetic preference. In fact, most traditional approaches adopting hand-crafted feature are, to some extent, able to understand part of image's aesthetic and content information while few studies have been conducted in the context of deep learning. Moreover, we discover that aesthetic rating is ambiguous so that many examples are uncertain in aesthetic level. This has caused a highly imbalanced distribution of aesthetic ratings. To tackle all these issues, we propose an end-to-end convolutional neural network (CNN) model which simultaneously implements aesthetic classification and understanding. To overcome the imbalanced aesthetic ratings, a sample-specific classification method that re-weights samples' importance is proposed. We find that dropping out ambiguous image, as common adopted by recent deep learning models, is a special case of the sample-specific method, and also figure out that as the weights of the non-ambiguous images increase, the performance is positively affected. In order to understand what is learned in the deep model, global average pooling (GAP) following the last feature map is employed to generate aesthetic activation map (AesAM) and attribute activation map (AttAM). AesAM and AttAM respectively represent the likelihood of aesthetic level for spatial location, and the likelihood of different attribute information. In particular, AesAM mainly accounts for what is learned in deep model. Experiments are carried out on public aesthetic datasets and state-of-the-art performance is achieved. Thanks to the introduction of AttAM, the aesthetic preferen

关键词： Iterative selective TrAdaBoost   modified correction factor   modified particle swarm optimization   optimal feature subspace   rapid convergence speed   scene classification   TrAdaBoost   weight mismatch  

摘要： Scene classification is usually based on supervised learning, but collecting instances is expensive and time-consuming. TrAdaBoost has achieved great success in transferring source instances to target images, but it has problems, such as the excessive focus on instances harder to classify, the rapid convergence speed of the source instances, and the weight mismatch caused by the big gap between the number of source and target instances, leading to decreased classification accuracy. In this paper, in order to address these problems, classical particle swarm optimization (PSO) is modified to select the optimal feature subspace for classifying the "harder" and "easier" instances by reducing unimportant dimensions. A modified correction factor is proposed by considering the classification accuracy of the instances from both domains, to decrease the convergence speed. Iterative selective TrAdaBoost is also proposed to reduce the weight mismatch by removing the indiscriminate source instances. The experimental results obtained with three benchmark data sets confirm that the proposed method outperforms most of the previous methods of scene classification with limited target samples. It is also proved that modified PSO for optimal feature subspace selection, the modified correction factor, and iterative selective TrAdaBoost are all effective in improving the classification accuracy, giving improvements of 3.6%, 4.3%, and 2.7%, and these three contributions together increase the classification accuracy by about 8% in total.

关键词： remote sensing image   convolutional neural network   optimal parameter   lack of sample   tensor analysis  

摘要： Recent research has shown that spatial-spectral information can help to improve the classification of hyperspectral images (HSIs). Therefore, three-dimensional convolutional neural networks (3D-CNNs) have been applied to HSI classification. However, a lack of HSI training samples restricts the performance of 3D-CNNs. To solve this problem and improve the classification, an improved method based on 3D-CNNs combined with parameter optimization, transfer learning, and virtual samples is proposed in this paper. Firstly, to optimize the network performance, the parameters of the 3D-CNN of the HSI to be classified (target data) are adjusted according to the single variable principle. Secondly, in order to relieve the problem caused by insufficient samples, the weights in the bottom layers of the parameter-optimized 3D-CNN of the target data can be transferred from another well trained 3D-CNN by a HSI (source data) with enough samples and the same feature space as the target data. Then, some virtual samples can be generated from the original samples of the target data to further alleviate the lack of HSI training samples. Finally, the parameter-optimized 3D-CNN with transfer learning can be trained by the training samples consisting of the virtual and the original samples. Experimental results on real-world hyperspectral satellite images have shown that the proposed method has great potential prospects in HSI classification.

关键词： Electronic tongue   Sensor array   Pharmaceuticals   Pattern recognition techniques   Taste-masking  

摘要： This work reports a critical evaluation of performance of various pattern recognition techniques applied to the classification of pharmaceutical taste-masked samples. Data obtained by potentiometric electronic tongue equipped with 16 ion-selective electrodes (ISEs) were processed by the most frequently used techniques in the analysis of electronic tongue data. Principal component analysis, partial least squares discriminant analysis, soft independent modelling of class analogy, principal component regression, support vector machine discriminant analysis, 3-way partial least squares, K-nearest neighbours as well as combination of principal components analysis and back propagation neural networks were tested. In order to compare their ability to estimate class affinity of pharmaceutical samples, sensitivity, precision, percent of correct classification (%cc) and root mean square error (RMSE) were calculated. Additionally, 4 different kinds of data matrices: dynamic responses, stationary responses, combinations of them both, CPA values (change of the membrane potential caused by adsorption) were processed by pattern recognition techniques for the determination of the influence of the extraction of the data on the classification results. SVM-DA is proved to exhibit the best performance for the most commonly applied data extraction i.e. the steady-state response of the sensor array. Furthermore, it is shown, that including dynamic responses in the data matrix better classification abilities of the majority of the studied pattern recognition techniques are obtained. It must be underlined, that the presented findings are based on studying 399 models for whom all performance factors (sensitivity, precision, %cc, RMSE) were determined for both train and test sets to obtain reliable and repeatable results. (C) 2018 Elsevier B.V. All rights reserved.