基于多核学习的多特征融合图像分类研究

图像分类任务是计算机视觉中的一个重要研究方向。组合多种特征在一定程度上能够使得图像分类准确度得到提高。然而,如何组合多种图像特征是一个悬而未决的难题。提出了一种基于多类多核学习的多特征融合算法,并应用到图像分类任务。算法在有效地利用多核学习自动选取对当前任务有价值特征的优势的同时,避免了在多核学习中将多类问题分解为多个二分问题。在图像特征表示方面,使用字典自学习方法。实验结果表明,提出的算法能够有效地提高图像分类的准确度。     

融合深度特征和多核增强学习的显著目标检测

目的 针对现有基于手工特征的显著目标检测算法对于显著性物体尺寸较大、背景杂乱以及多显著目标的复杂图像尚不能有效抑制无关背景区域且完整均匀高亮显著目标的问题,提出了一种利用深度语义信息和多核增强学习的显著目标检测算法。方法 首先对输入图像进行多尺度超像素分割计算,利用基于流形排序的算法构建弱显著性图。其次,利用已训练的经典卷积神经网络对多尺度序列图像提取蕴含语义信息的深度特征,结合弱显著性图从多尺度序列图像内获得可靠的训练样本集合,采用多核增强学习方法得到强显著性检测模型。然后,将该强显著性检测模型应用于多尺度序列图像的所有测试样本中,线性加权融合多尺度的检测结果得到区域级的强显著性图。最后,根据像素间的位置和颜色信息对强显著性图进行像素级的更新,以进一步提高显著图的准确性。结果 在常用的MSRA5K、ECSSD和SOD数据集上与9种主流且相关的算法就准确率、查全率、F-measure值、准确率—召回率（PR）曲线、加权F-measure值和覆盖率（OR）值等指标和直观的视觉检测效果进行了比较。相较于性能第2的非端到端深度神经网络模型,本文算法在3个数据集上的平均F-measure值、加权F-measure值、OR值和平均误差（MAE）值,分别提高了1.6%,22.1%,5.6%和22.9%。结论 相较于基于手工特征的显著性检测算法,本文算法利用图像蕴含的语义信息并结合多个单核支持向量机（SVM）分类器组成强分类器,在复杂图像上取得了较好的检测效果。     

A primal method for multiple kernel learning

The canonical support vector machines (SVMs) are based on a single kernel, recent publications have shown that using multiple kernels instead of a single one can enhance interpretability of the decision function and promote classification accuracy. However, most of existing approaches mainly reformulate the multiple kernel learning as a saddle point optimization problem which concentrates on solving the dual. In this paper, we show that the multiple kernel learning (MKL) problem can be reformulated as a BiConvex optimization and can also be solved in the primal. While the saddle point method still lacks convergence results, our proposed method exhibits strong optimization convergence properties. To solve the MKL problem, a two-stage algorithm that optimizes canonical SVMs and kernel weights alternately is proposed. Since standard Newton and gradient methods are too time-consuming, we employ the truncated-Newton method to optimize the canonical SVMs. The Hessian matrix need not be stored explicitly, and the Newton direction can be computed using several Preconditioned Conjugate Gradient steps on the Hessian operator equation, the algorithm is shown more efficient than the current primal approaches in this MKL setting. Furthermore, we use the Nesterov’s optimal gradient method to optimize the kernel weights. One remarkable advantage of solving in the primal is that it achieves much faster convergence rate than solving in the dual and does not require a two-stage algorithm even for the single kernel LapSVM. Introducing the Laplacian regularizer, we also extend our primal method to semi-supervised scenario. Extensive experiments on some UCI benchmarks have shown that the proposed algorithm converges rapidly and achieves competitive accuracy.     

多核学习纹理特征的立体图像质量评价

为了有效地评价各种失真类型双目立体图像的质量,提出利用多核学习机学习立体图像平面纹理信息和3D映射信息的通用无参考立体图像质量评价IQA方法。该方法首先利用立体匹配模型对左右视图进行处理,获得相应的视差图DM和误差能量图DMEE;对左右视图、视差图和误差能量图进行相位一致性和结构张量变换,获得它们的平坦区和边缘区;分别提取左右视图两个区域纹理特征作为平面信息,提取视差图的纹理特征和误差能量图的统计特征作为3D信息;将所有特征作为多核学习机的输入,利用多核学习的信息融合能力预测待测失真立体图像质量。由于充分利用了立体图像的左右视图、视差图和误差能量图的失真信息,以及多核学习的信息融合能力,该方法具有很好的前景。在LIVE 3D图像质量数据库上的实验表明,该方法与主观质量有较高一致性,与现有的双目立体质量评价方法相比有很大的竞争力。     

Identification of human microRNA-disease association via low-rank approximation-based link propagation and multiple kernel learning

Numerous studies have demonstrated that human microRNAs (miRNAs) and diseases are associated and studies on the microRNA-disease association (MDA) have been conducted. We developed a model using a low-rank approximation-based link propagation algorithm with Hilbert–Schmidt independence criterion-based multiple kernel learning (HSIC-MKL) to solve the problem of the large time commitment and cost of traditional biological experiments involving miRNAs and diseases, and improve the model effect. We constructed three kernels in miRNA and disease space and conducted kernel fusion using HSIC-MKL. Link propagation uses matrix factorization and matrix approximation to effectively reduce computation and time costs. The results of the experiment show that the approach we proposed has a good effect, and, in some respects, exceeds what existing models can do.     

Ensemble learning from multiple information sources via label propagation and consensus

Many applications are facing the problem of learning from multiple information sources, where sources may be labeled or unlabeled, and information from multiple information sources may be beneficial but cannot be integrated into a single information source for learning. In this paper, we propose an ensemble learning method for different labeled and unlabeled sources. We first present two label propagation methods to infer the labels of training objects from unlabeled sources by making a full use of class label information from labeled sources and internal structure information from unlabeled sources, which are processes referred to as global consensus and local consensus, respectively. We then predict the labels of testing objects using the ensemble learning model of multiple information sources. Experimental results show that our method outperforms two baseline methods. Meanwhile, our method is more scalable for large information sources and is more robust for labeled sources with noisy data.     

An information theoretic sparse kernel algorithm for online learning

Kernel-based algorithms have been proven successful in many nonlinear modeling applications. However, the computational complexity of classical kernel-based methods grows superlinearly with the increasing number of training data, which is too expensive for online applications. In order to solve this problem, the paper presents an information theoretic method to train a sparse version of kernel learning algorithm. A concept named instantaneous mutual information is investigated to measure the system reliability of the estimated output. This measure is used as a criterion to determine the novelty of the training sample and informative ones are selected to form a compact dictionary to represent the whole data. Furthermore, we propose a robust learning scheme for the training of the kernel learning algorithm with an adaptive learning rate. This ensures the convergence of the learning algorithm and makes it converge to the steady state faster. We illustrate the performance of our proposed algorithm and compare it with some recent kernel algorithms by several experiments.     

Analytic center cutting plane method for multiple kernel learning

Multiple Kernel Learning (MKL) is a popular generalization of kernel methods which allows the practitioner to optimize over convex combinations of kernels. We observe that many recent MKL solutions can be cast in the framework of oracle based optimization, and show that they vary in terms of query point generation. The popularity of such methods is because the oracle can fortuitously be implemented as a support vector machine. Motivated by the success of centering approaches in interior point methods, we propose a new approach to optimize the MKL objective based on the analytic center cutting plane method (accpm). Our experimental results show that accpm outperforms state of the art in terms of rate of convergence and robustness. Further analysis sheds some light as to why MKL may not always improve classification accuracy over naive solutions.     

A framework for integrating information sources under lattice structure

Different observations over one and the same natural phenomenon often lead to different collections of information elements describing that phenomenon. Such collections of information elements, which we call information sources, can be heterogeneous, redundant, complementary, or even contradictory. In general, the information elements describing the phenomenon can be unified and related in a hierarchical structure, using existing methods for data integration and ontology merging. In this paper, we consider the problem of redundancy, complementarity, and consistency of information sources, under the assumption that the information elements of each source are related in a lattice structure. We propose various methods for integrating information sources and establish relationships between these methods. Applications of the framework are illustrated through examples in the areas of geographic information and battlefield target identification.     

Human activity recognition using multi-features and multiple kernel learning

This paper presents two sets of features, shape representation and kinematic structure, for human activity recognition using a sequence of RGB-D images. The shape features are extracted using the depth information in the frequency domain via spherical harmonics representation. The other features include the motion of the 3D joint positions (i.e. the end points of the distal limb segments) in the human body. Both sets of features are fused using the Multiple Kernel Learning (MKL) technique at the kernel level for human activity recognition. Our experiments on three publicly available datasets demonstrate that the proposed features are robust for human activity recognition and particularly when there are similarities among the actions.     

Bridging deep and multiple kernel learning: A review

Kernel methods and deep learning are two of the most currently remarkable machine learning techniques that have achieved great success in many applications. Kernel methods are powerful tools to capture nonlinear patterns behind data. They implicitly learn high (even infinite) dimensional nonlinear features in the reproducing kernel Hilbert space (RKHS) while making the computation tractable by leveraging the kernel trick. It is commonly agreed that the success of kernel methods is very much dependent on the choice of kernel. Multiple kernel learning (MKL) is one possible scheme that performs kernel combination and selection for a variety of learning tasks, such as classification, clustering, and dimensionality reduction. Deep learning models project input data through several layers of nonlinearity and learn different levels of abstraction. The composition of multiple layers of nonlinear functions can approximate a rich set of naturally occurring input-output dependencies. To bridge kernel methods and deep learning, deep kernel learning has been proven to be an effective method to learn complex feature representations by combining the nonparametric flexibility of kernel methods with the structural properties of deep learning. This article presents a comprehensive overview of the state-of-the-art approaches that bridge the MKL and deep learning techniques. Specifically, we systematically review the typical hybrid models, training techniques, and their theoretical and practical benefits, followed by remaining challenges and future directions. We hope that our perspectives and discussions serve as valuable references for new practitioners and theoreticians seeking to innovate in the applications of the approaches incorporating the advantages of both paradigms and exploring new synergies.     

RGBD co-saliency detection via multiple kernel boosting and fusion

Multimedia Tools and Applications - RGBD co-saliency detection, which aims at extracting common salient objects from a group of RGBD images with the additional depth information, has become an...     

Personalized rough-set-based recommendation by integrating multiple contents and collaborative information

In recent years, explosively-growing information makes the users confused in making decisions among various kinds of products such as music, movies, books, etc. As a result, it is a challenging issue to help the user identify what she/he prefers. To this end, so called recommender systems are proposed to discover the implicit interests in user’s mind based on the usage logs. However, the existing recommender systems suffer from the problems of cold-start, first-rater, sparsity and scalability. To alleviate such problems, we propose a novel recommender, namely FRSA (Fusion of Rough-Set and Average-category-rating) that integrates multiple contents and collaborative information to predict user’s preferences based on the fusion of Rough-Set and Average-category-rating. Through the integrated mining of multiple contents and collaborative information, our proposed recommendation method can successfully reduce the gap between the user’s preferences and the automated recommendations. The empirical evaluations reveal that the proposed method, FRSA, can associate the recommended items with user’s interests more effectively than other existing well-known ones in terms of accuracy.     

Human action recognition with salient trajectories and multiple kernel learning

Action recognition in videos plays an important role in the field of computer vision and multimedia, and there exist lots of challenges due to the complexity of spatial and temporal information. Trajectory-based approach has shown to be efficient recently, and a new framework and algorithm of trajectory space information based multiple kernel learning (TSI-MKL) is exploited in this paper. First, dense trajectories are extracted as raw features, and three saliency maps are computed corresponding to color, space, and optical flow on frames at the same time. Secondly, a new method combining above saliency maps is proposed to filter the achieved trajectories, by which a set of salient trajectories only containing foreground motion regions is obtained. Afterwards, a novel two-layer clustering is developed to cluster the obtained trajectories into several semantic groups and the ultimate video representation is generated by encoding each group. Finally, representations of different semantic groups are fed into the proposed kernel function of a multiple kernel classifier. Experiments are conducted on three popular video action datasets and the results demonstrate that our presented approach performs competitively compared with the state-of-the-art.     

Linear feature extraction by integrating pairwise and global discriminatory information via sequential forward floating selection and kernel QR factorization with column pivoting

Linear discriminant analysis (LDA) is often used to produce an effective linear feature extractor for classification. However, some approaches of LDA, such as Fisher's linear discriminant, are not robust to outlier classes. In this paper, a novel approach is proposed to robustly produce an effective linear feature extractor by integrating the discriminatory information from the global and pairwise approaches of LDA. The discriminatory information is integrated either by the sequential forward floating selection algorithm with a criterion function based on the Chernoff bound or by ranking the discriminatory information using the kernel QR factorization with column pivoting according to the indication of an applicability index for these two methods. The proposed approach was compared to various methods of LDA. The experimental results have shown the robustness of the proposed approach and proved the feasibility of the proposed approach.     

利用多核增强学习的立体图像舒适度评价模型

目的 传统的立体视觉舒适度评价模型,在学习阶段一般采用回归算法,且需要大量的包含主观测试数据的训练样本,针对这个问题,提出一种利用多核增强学习分类算法的立体图像舒适度评价模型。方法 首先,考虑人们在实际观测图像时,对于先后观测到的不同图像进行相互比较的情况,将评价模型看成是偏好分类器,构造包含偏好标签的偏好立体图像对（PSIP）,构成PSIP训练集;其次,提取多个视差统计特征和神经学模型响应特征;然后,利用基于AdaBoost的多核学习算法来建立偏好标签与特征之间的关系模型,并分析偏好分类概率（即相对舒适度概率）与最终的视觉舒适度之间的映射关系。结果 在独立立体图像库上,与现有代表性回归算法相比较,本文算法的Pearson线性相关系数（PLCC）在0.84以上,Spearman等级相关系数（SRCC）在0.80以上,均优于其他模型的各评价指标;而在跨库测试中,本文算法的PLCC、SRCC指标均优于传统的支持向量回归算法。结论 相比于传统的回归算法,本文算法具有更好的评价性能,能够更为准确地预测立体图像视觉舒适度。     

融合分层分块信息的轧制过程运行状态评估方法及应用

工业过程运行状态评估方法是对过程当前运行状态进行合理的评价,为工业过程的安全、高效运行提供有益的指导.带钢轧制过程具有流程长且系统层级多等特点,而传统的运行状态评估往往采用集中式的评估方法,难以对轧制过程全流程的运行状态进行合理的评估.针对此问题,提出一种融合分层分块信息的轧制过程运行状态评估方法.采用一种多层级分块的评估策略,将全流程分为若干个层级和子块,提高评估结果的可解释性.提出一种联合核主成分分析(kernel principal component analysis,KPCA)和t-分布随机邻域嵌入算法(t-distributed stochastic neighbor embedding,t-SNE)的特征提取方法,并行地提取全局和局部的特征信息.进而,针对传统支持向量机(support vector machine,SVM)输出结果为硬判型输出,将SVM的输出结果映射为后验概率,并通过D-S证据理论融合多个层级的运行状态评估结果,从而实现决策层面的信息融合,提高评估结果的准确性.最后,将所提出方法应用于实际带钢轧制过程,与各类传统的方法相比评估准确率提高近18%.     

Local weighting schemes for protein multiple sequence alignment

This paper describes three weighting schemes for improving the accuracy of progressive multiple sequence alignment methods: (1) global profile pre-processing, to capture for each sequence information about other sequences in a profile before the actual multiple alignment takes place; (2) local pre-processing; which incorporates a new protocol to only use non-overlapping local sequence regions to construct the pre-processed profiles; and (3) local-global alignment, a weighting scheme based on the double dynamic programming (DDP) technique to softly bias global alignment to local sequence motifs. The first two schemes allow the compilation of residue-specific multiple alignment reliability indices, which can be used in an iterative fashion. The schemes have been implemented with associated iterative modes in the PRALINE multiple sequence alignment method, and have been evaluated using the BAliBASE benchmark alignment database. These tests indicate that PRALINE is a toolbox able to build alignments with very high quality. We found that local profile pre-processing raises the alignment quality by 5.5% compared to PRALINE alignments generated under default conditions. Iteration enhances the quality by a further percentage point. The implications of multiple alignment scoring functions and iteration in relation to alignment quality and benchmarking are discussed.