基于残差神经网络的图像超分辨率改进算法

为更有效地提升图像的超分辨率（SR）效果,提出了一种多阶段级联残差卷积神经网络模型。首先,该模型采用了两阶段超分辨率图像重建方法先重建2倍超分辨率图像,再重建4倍超分辨率图像;其次,第一阶段与第二阶段皆使用残差层和跳层结构预测出高分辨率空间的纹理信息,由反卷积层分别重建出2倍与4倍大小的超分辨率图像;最后,以两阶段的结果分别构建多任务损失函数,利用第一阶段的损失指导第二阶段的损失,从而提高网络的训练速度,加强网络学习中的监督指导。实验结果表明,与bilinear算法、bicubic算法、基于卷积神经网络的图像超分辨率（SRCNN）算法和加速的超分辨率卷积神经网络（FSRCNN）算法相比,所提模型能更好地重建出图像的细节和纹理,避免了经过迭代之后造成的图像过度平滑,获得更高的峰值信噪比（PSNR）和平均结构相似度（MSSIM）。     

基于多粒度特征生成对抗网络的跨分辨率行人重识别

现有基于生成对抗网络（GAN）的超分辨率（SR）重建方法用于跨分辨率行人重识别（ReID）时,重建图像在纹理结构内容的恢复和特征一致性保持方面均存在不足。针对上述问题,提出基于多粒度信息生成网络的跨分辨率行人ReID方法。首先,在生成器的多层网络上均引入自注意力机制,聚焦多粒度稳定的结构关联区域,重点恢复低分辨率（LR）行人图像的纹理结构信息;同时,在生成器后增加一个识别器,在训练过程中最小化生成图像与真实图像在不同粒度特征上的损失,提升生成图像与真实图像在特征上的一致性。然后,联合自注意力生成器和识别器,与判别器交替优化,在内容和特征上改进生成图像。最后,联合改进的GAN和行人ReID网络交替训练优化网络的模型参数,直至模型收敛。在多个跨分辨率行人数据集上的实验结果表明,所提算法的累计匹配曲线（CMC）在其首选识别率（rank?1）上的准确率较现有同类算法平均提升10个百分点,在提升SR图像内容一致性和特征表达一致性方面均表现更优。     

Event prediction algorithm using neural networks for the power management system of electric vehicles

The power management system for electronic vehicles selectively activates Electronic Control Units (ECUs) in the electronic control system according to time-series vehicle data and predefined operation states. However, at an operation state transition, the energy overheads used for the selective ECU activation could be higher than the energy saved by deactivating ECUs. To prevent these energy-inefficient state transitions, we apply two main ideas to our proposed algorithm: (A) unacceptable state transitions and (B) adaptive training speed. For the unacceptable transitions, our energy model evaluates the breakeven time where energy saving equals to energy overheads. Based on the breakeven time, our algorithm classifies training dataset as unacceptable and acceptable event sets. Especially when the algorithm trains neural networks for the two event sets, the adaptive training speed expedites its training speed based on a history of training errors. Consequently, without violating in-vehicle time constraints, the algorithm could provide real-time predictions and save energy overheads by avoiding unacceptable transitions. In the simulation results on real driving datasets, our algorithm improves the energy dissipation of the electronic control system by 5% to 7%.     

基于保局部核RVM的说话人识别方法

针对说话人语音特征随音量、情绪、健康等因素变化呈现出的复杂分布结构,提出一种基于保局部核相关向量机(RVM)的说话人识别方法。在RVM模型所采用的高斯核函数中引入相似度因子,以保留数据局部结构,构成保局部核RVM模型。在模型训练过程中采用快速算法以避免大型矩阵逆操作,减少计算量,可适用于大样本场合。应用结果表明,该方法能加快测试速度,提高分类精度。     

Handling numeric attributes when comparing Bayesian network classifiers: does the discretization method matter?

Within the framework of Bayesian networks (BNs), most classifiers assume that the variables involved are of a discrete nature, but this assumption rarely holds in real problems. Despite the loss of information discretization entails, it is a direct easy-to-use mechanism that can offer some benefits: sometimes discretization improves the run time for certain algorithms; it provides a reduction in the value set and then a reduction in the noise which might be present in the data; in other cases, there are some Bayesian methods that can only deal with discrete variables. Hence, even though there are many ways to deal with continuous variables other than discretization, it is still commonly used. This paper presents a study of the impact of using different discretization strategies on a set of representative BN classifiers, with a significant sample consisting of 26 datasets. For this comparison, we have chosen Naive Bayes (NB) together with several other semi-Naive Bayes classifiers: Tree-Augmented Naive Bayes (TAN), k-Dependence Bayesian (KDB), Aggregating One-Dependence Estimators (AODE) and Hybrid AODE (HAODE). Also, we have included an augmented Bayesian network created by using a hill climbing algorithm (BNHC). With this comparison we analyse to what extent the type of discretization method affects classifier performance in terms of accuracy and bias-variance discretization. Our main conclusion is that even if a discretization method produces different results for a particular dataset, it does not really have an effect when classifiers are being compared. That is, given a set of datasets, accuracy values might vary but the classifier ranking is generally maintained. This is a very useful outcome, assuming that the type of discretization applied is not decisive future experiments can be d times faster, d being the number of discretization methods considered.     

基于半耦合稀疏表达的极低分辨率人脸识别

现有基于学习的人脸超分辨率算法假设高低分辨率特征具有流形一致性（耦合字典学习）,然而低分辨率图像的降质过程使得高低分辨率特征产生了“一对多”的映射关系偏差,减少了极低分辨率图像特征的判决信息,降低了超分辨率重建图像的识别率。针对这一问题,引入了半耦合稀疏字典学习模型,松弛高低分辨率流形一致性假设,同时学习稀疏表达字典和稀疏表达系数之间的映射函数,提升高低分辨率判决特征的一致性,在此基础上,引入协同分类模型,实现半耦合特征的高效分类。实验表明：相比于传统稀疏表达分类算法,算法不仅提高了识别率,并且还大幅度降低了时间开销,验证了半耦合稀疏学习字典在人脸识别中的有效性。     

在逐渐缩小的空间上渐进学习朴素贝叶斯参数

局部加权朴素贝叶斯(LWNB)是朴素贝叶斯(NB)的一种较好的改进,判别频率估计(DFE)可以极大地提高NB的泛化正确率。受LWNB和DFE启发,提出逐渐缩小空间(GCS)算法用来学习NB参数：对于一个测试实例,寻找包含全体训练实例的全局空间的一系列逐渐缩小的子空间。这些子空间具有两种性质：1)它们都包含测试实例;2)一个空间一定包含在任何一个比它大的空间中。在逐渐缩小的空间上使用修改的DFE(MDFE)算法渐进地学习NB的参数,然后使用NB分类测试实例。与LWNB的根本不同是：GCS使用全体训练实例学习NB并且GCS可以实现为非懒惰版本。实现了GCS的决策树版本(GCS-T),实验结果显示,与C4.5以及贝叶斯分类算法(如Naive Bayes、BaysianNet、NBTree、LWNB、隐朴素贝叶斯)相比,GCS-T具有较高的泛化正确率,并且GCS-T的分类速度明显快于LWNB。     

联合多任务学习的人脸超分辨率重建

目的 人脸超分辨率重建是特定应用领域的超分辨率问题,为了充分利用面部先验知识,提出一种基于多任务联合学习的深度人脸超分辨率重建算法。方法 首先使用残差学习和对称式跨层连接网络提取低分辨率人脸的多层次特征,根据不同任务的学习难易程度设置损失权重和损失阈值,对网络进行多属性联合学习训练。然后使用感知损失函数衡量HR（high-resolution）图像与SR（super-resolution）图像在语义层面的差距,并论证感知损失在提高人脸语义信息重建效果方面的有效性。最后对人脸属性数据集进行增强,在此基础上进行联合多任务学习,以获得视觉感知效果更加真实的超分辨率结果。结果 使用峰值信噪比（PSNR）和结构相似度（SSIM）两个客观评价标准对实验结果进行评价,并与其他主流方法进行对比。实验结果显示,在人脸属性数据集（CelebA）上,在放大8倍时,与通用超分辨率MemNet（persistent memory network）算法和人脸超分辨率FSRNet（end-to-end learning face super-resolution network）算法相比,本文算法的PSNR分别提升约2.15 dB和1.2 dB。结论 实验数据与效果图表明本文算法可以更好地利用人脸先验知识,产生在视觉感知上更加真实和清晰的人脸边缘和纹理细节。     

像素归一化方法在恶意代码可视分析中的应用

恶意代码的编写者通常采用自动化的手段开发恶意代码变种,使得恶意代码的数量呈现迅猛增长的态势。由于自动化的方式会重复利用恶意代码中的核心模块,因此也为病毒研究人员辨识和区分恶意代码族提供了有利依据。借鉴灰度图的思想,利用K-Nearest Neighbor（KNN）分类算法,给出了一种新的研究恶意代码谱系分类的可视化方法。其基本思想是,通过将二进制文件转换成双色通道的位图和像素归一图,从可视化的角度标识恶意样本特性,以此实现恶意代码族的相似度比较及分类。实验结果表明采用了像素归一化的降维映射机制能显著地减小文件可视特征的呈现时间开销,且该方法以自动化操作的方式运用Jaccard距离算法进行快速相似度比较,实现了恶意代码样本的有效分类,提高了分析人员的识别效率。     

Extended Spectral Regression for efficient scene recognition

This paper proposes a novel method based on Spectral Regression (SR) for efficient scene recognition. First, a new SR approach, called Extended Spectral Regression (ESR), is proposed to perform manifold learning on a huge number of data samples. Then, an efficient Bag-of-Words (BOW) based method is developed which employs ESR to encapsulate local visual features with their semantic, spatial, scale, and orientation information for scene recognition. In many applications, such as image classification and multimedia analysis, there are a huge number of low-level feature samples in a training set. It prohibits direct application of SR to perform manifold learning on such dataset. In ESR, we first group the samples into tiny clusters, and then devise an approach to reduce the size of the similarity matrix for graph learning. In this way, the subspace learning on graph Laplacian for a vast dataset is computationally feasible on a personal computer. In the ESR-based scene recognition, we first propose an enhanced low-level feature representation which combines the scale, orientation, spatial position, and local appearance of a local feature. Then, ESR is applied to embed enhanced low-level image features. The ESR-based feature embedding not only generates a low dimension feature representation but also integrates various aspects of low-level features into the compact representation. The bag-of-words is then generated from the embedded features for image classification. The comparative experiments on open benchmark datasets for scene recognition demonstrate that the proposed method outperforms baseline approaches. It is suitable for real-time applications on mobile platforms, e.g. tablets and smart phones.     

Naive Bayes for value difference metric

The value difference metric (VDM) is one of the best-known and widely used distance functions for nominal attributes. This work applies the instanceweighting technique to improveVDM. An instance weighted value difference metric (IWVDM) is proposed here. Different from prior work, IWVDM uses naive Bayes (NB) to find weights for training instances. Because early work has shown that there is a close relationship between VDM and NB, some work on NB can be applied to VDM. The weight of a training instance x, that belongs to the class c, is assigned according to the difference between the estimated conditional probability ^P(c|x) by NB and the true conditional probability P(c|x), and the weight is adjusted iteratively. Compared with previous work, IWVDM has the advantage of reducing the time complexity of the process of finding weights, and simultaneously improving the performance of VDM. Experimental results on 36 UCI datasets validate the effectiveness of IWVDM.     

Infrared Image Super-Resolution with Parallel Random Forest

Infrared imaging has the advantage of all-weather working ability. Due to the limitation of the hardware and the high cost, the resolution of infrared image (IR) is very low. To improve the resolution of IR images, this paper exploits super-resolution (SR) method for IR images. A new SR framework by using random forests is proposed in this paper. Existing methods adopts single regression model for SR. However, which single regression model tends to overfit training data, and would lead to a poor performance. Furthermore, the existing methods are not suitable for real-time system due to the heavy time consuming. To resolve this problem, an ensemble regression model, i.e. random forests rather than single regression model is adopted in this paper. In addition, to achieve better results multi-regression models rather than a single regression model are trained on the clustered training data. Moreover, the features used in many SR methods cannot extract features on diagonal orientation. To resolve this problem, we adopt a second order derivative filter, which can extract features on diagonal orientation. The experimental results demonstrate the availability of the proposed method.     

联合RMSE损失LSTM-CNN模型的股价预测

由于金融市场的复杂性和多变性,当前模型尚不能完全覆盖股票走势影响因素的方方面面,在预测精度方面还存在改进空间.基于此,提出了一种联合RMSE损失LSTM-CNN(long short-term memory-convolutional neural networks)的方法.该方法创新性地通过联合两个模型的RMSE损失...     

Feature selection for the naive bayesian classifier using decision trees

It is known that Naive Bayesian classifier (NB) works very well on some domains, and poorly on others. The performance of NB suffers in domains that involve correlated features. C4.5 decision trees, on the other hand, typically perform better than the Naive Bayesian algorithm on such domains. This paper describes a Selective Bayesian classifier (SBC) that simply uses only those features that C4.5 would use in its decision tree when learning a small example of a training set, a combination of the two different natures of classifiers. Experiments conducted on ten data sets indicate that SBC performs markedly better than NB on all domains, and SBC outperforms C4.5 on many data sets of which C4.5 outperform NB. Augmented Bayesian classifier (ABC) is also tested on the same data, and SBC appears to perform as well as ABC. SBC also can eliminate, in most cases, more than half of the original attributes, which can greatly reduce the size of the training and test data as well as the running time. Further, the SBC algorithm typically learns faster than both C4.5 and NB, needing fewer training examples to reach a high accuracy of classifications.     

A hybrid multilevel-active set method for large box-constrained linear discrete ill-posed problems

Many questions in science and engineering give rise to linear ill-posed problems, whose solution is known to satisfy box constraints, such as nonnegativity. The solution of discretized versions of these problems is highly sensitive to perturbations in the data, discretization errors, and round-off errors introduced during the computations. It is therefore often beneficial to impose known constraints during the solution process. This paper describes a two-phase algorithm for the solution of large-scale box-constrained linear discrete ill-posed problems. The first phase applies a cascadic multilevel method and imposes the constraints on each level by orthogonal projection. The second phase improves the computed approximate solution on the finest level by an active set method. The latter allows several indices of the active set to be updated simultaneously. This reduces the computational effort significantly, when compared to standard active set methods that update one index at a time. Applications to image restoration are presented.     

基于生成式对抗网络的图像超分辨率重建算法

针对传统图像超分辨率重建算法存在网络训练困难与生成图像存在伪影的问题,提出一种利用生成式对抗网络的超分辨率重建算法。去除生成式对抗网络的批量归一化层降低计算复杂度,将其中的残差块替换为密集残差块构成生成网络,使用VGG19网络作为判别网络的基础框架,以全局平均池化代替全连接层防止过拟合,引入纹理损失函数、感知损失函数、对抗损失函数和内容损失函数构成生成器的总目标函数,利用纹理损失增强局部信息匹配度,采用激活层前的特征信息计算感知损失获取更多细节特征,使用WGAN-GP理论优化网络模型的对抗损失加速收敛,运用内容损失提升图像低频信息的准确性。实验结果表明,该算法重建图像的平均峰值信噪比为27.97 dB,平均结构相似性为0.777,与SRGAN和EDSR等算法相比,其在未延长较多运行时间的情况下,重建结果的纹理细节更清晰且亮度信息更准确,更符合视觉感官评价要求。     

Evaluation of the MODIS Albedo product over a heterogeneous agricultural area

In this article, the Moderate Resolution Imaging Spectroradiometer (MODIS) Bidirectional Reflectance Distribution Function (BRDF)/Albedo product (MCD43) is evaluated over a heterogeneous agricultural area in the framework of the Earth Observation: Optical Data Calibration and Information Extraction (EODIX) project campaign, which was developed in Barrax (Spain) in June 2011. In this method, two models, the RossThick-LiSparse-Reciprocal (RTLSR) (which corresponds to the MODIS BRDF algorithm) and the RossThick-Maignan-LiSparse-Reciprocal (RTLSR-HS), were tested over airborne data by processing high-resolution images acquired with the Airborne Hyperspectral Scanner (AHS) sensor. During the campaign, airborne images were retrieved with different view zenith angles along the principal and orthogonal planes. Comparing the results of applying the models to the airborne data with ground measurements, we obtained a root mean square error (RMSE) of 0.018 with both RTLSR and RTLSR-HS models. The evaluation of the MODIS BRDF/Albedo product (MCD43) was performed by comparing satellite images with AHS estimations. The results reported an RMSE of 0.04 with both models. Additionally, taking advantage of a homogeneous barley pixel, we compared in situ albedo data to satellite albedo data. In this case, the MODIS albedo estimation was (0.210 ± 0.003), while the in situ measurement was (0.204 ± 0.003). This result shows good agreement in regard to a homogeneous pixel.     

一种基于遗传算法的BP网络改进方法

为克服和改进传统的BP算法的不足，发挥神经网络和遗传算法各自的优势，提出一种新的基于遗传算法的改进的BP网络训练方法。在美国手写体数字标准数据集MNIST库的实验结果表明，该方法提高了识别率，增加了网络的泛化能力，并且极大地节省了存储空间，缩短了学习时间。     

基于结构自相似性和形变块特征的单幅图像超分辨率算法

针对单幅图像超分辨率（SR）复原样本资源不足和抗噪性差的问题，提出一种基于结构自相似和形变块特征的单幅图像超分辨率算法。首先，该方法通过构建尺度模型，尽可能地扩展搜索空间，克服单幅图像超分辨率训练样本不足的缺陷；接着，通过样例块的几何形变提升了局限性的内部字典大小；最后，为了提升重建图片的抗噪性，利用组稀疏学习字典来重建图像。实验结果表明：与Bicubic、稀疏字典学习（ScSR）算法和基于卷积神经网络的超分辨率（SRCNN）等优秀字典学习算法相比，所提算法可以得到主观视觉效果更为清晰和客观评价更高的超分辨率图像，峰值信噪比（PSNR）平均约提升了0.35 dB。另外所提算法通过几何形变的方式扩展了字典规模和搜索的准确性，在算法时间消耗上平均约减少了80 s。     

Fractional snow cover mapping through artificial neural network analysis of MODIS surface reflectance

Accurate areal measurements of snow cover extent are important for hydrological and climate modeling. The traditional method of mapping snow cover is binary where a pixel is considered either snow-covered or snow-free. Fractional snow cover (FSC) mapping can achieve a more precise estimate of areal snow cover extent by estimating the fraction of a pixel that is snow-covered. The most common snow fraction methods applied to Moderate Resolution Imaging Spectroradiometer (MODIS) images have been spectral unmixing and an empirical Normalized Difference Snow Index (NDSI). Machine learning is an alternative for estimating FSC as artificial neural networks (ANNs) have been successfully used for estimating the subpixel abundances of other surfaces. The advantages of ANNs are that they can easily incorporate auxiliary information such as land cover type and are capable of learning nonlinear relationships between surface reflectance and snow fraction. ANNs are especially applicable to mapping snow cover extent in forested areas where spatial mixing of surface components is nonlinear. This study developed a multilayer feed-forward ANN trained through backpropagation to estimate FSC using MODIS surface reflectance, NDSI, Normalized Difference Vegetation Index (NDVI) and land cover as inputs. The ANN was trained and validated with higher spatial-resolution FSC maps derived from Landsat Enhanced Thematic Mapper Plus (ETM+) binary snow cover maps. Testing of the network was accomplished over training and independent test areas. The developed network performed adequately with RMSE of 12% over training areas and slightly less accurately over the independent test scenes with RMSE of 14%. The developed ANN also compared favorably to the standard MODIS FSC product. The study also presents a comprehensive validation of the standard MODIS snow fraction product whose performance was found to be similar to that of the ANN.