应用于不完整流形的ISOMAP算法

等距特征映射(ISOMAP)算法要求数据位于单一流形之上且具有良好采样,而当数据采样于一个不完整流形时,该算法将会产生“过聚类”问题。为此,提出了一种改进算法--WISOMAP,它采用多维尺度分析(MDS)算法的一个变种--WMDS来降低逼近精度相对较差的多边测地距离在MDS距离保持中的主导作用,使逼近精度相对较好的少边测地距离能够得到更好的保持,从而能在一定程度上缓解“过聚类”问题。实验结果表明WISOMAP算法能更好地对采样于不完整流形的数据进行可视化。     

cTraj: efficient indexing and searching of sequences containing multiple moving objects

Indexing sequences containing multiple moving objects by all features of these objects captured at every clock tick results in huge index structures due to the large number of extracted features in all sampled instances. Thus, the main problems with current systems that index sequences containing multiple moving objects are: huge storage requirements for index structures, slow search time and low accuracy due to lack of representation of the time-varying features of objects. In this paper, a technique called cTraj to address these problems is proposed. For each object in a sequence, cTraj captures the features at sampled instances. Then, it maps the object??s features at each sampled instance from high-dimensional feature space into a point in low-dimensional distance space. The sequence of points of an object in low-dimensional space is considered the time-varying feature trajectory of the object. To reduce storage requirements of an index structure, the sequence of points in each trajectory is represented by a minimum bounding box (MBB). cTraj indexes a sequence by the MBBs of its objects using a spatial access method (SAM), such as an R?tree; thus, greatly reducing storage requirements of the index and speeding up the search time. The cTraj technique does not result in any false dismissal, but the result might contain a few false alarms, which are removed by a two-step refinement process. The experiments show that the proposed cTraj technique produces effective results comparable to those of a sequential method, however much more efficient.     

引入外部词向量的文本信息网络表示学习

针对信息网络（text-based information network）现有研究多基于网络自身信息建模,受限于任务语料规模,只使用任务相关文本进行建模容易产生语义漂移或语义残缺的问题,本文将外部语料引入建模过程中,利用外部语料得到的词向量对建模过程进行优化,提出基于外部词向量的网络表示模型NE-EWV（network embedding based on external word vectors）,从语义特征空间以及结构特征空间两个角度学习特征融合的网络表示。通过实验,在现实网络数据集中对模型有效性进行了验证。实验结果表明,在链接预测任务中的AUC指标,相比只考虑结构特征的模型提升7%~19%,相比考虑结构与文本特征的模型在大部分情况下有1%~12%提升;在节点分类任务中,与基线方法中性能最好的CANE性能相当。证明引入外部词向量作为外部知识能够有效提升网络表示能力。     

Parametric embedding for class visualization

We propose a new method, parametric embedding (PE), that embeds objects with the class structure into a low-dimensional visualization space. PE takes as input a set of class conditional probabilities for given data points and tries to preserve the structure in an embedding space by minimizing a sum of Kullback-Leibler divergences, under the assumption that samples are generated by a gaussian mixture with equal covariances in the embedding space. PE has many potential uses depending on the source of the input data, providing insight into the classifier's behavior in supervised, semisupervised, and unsupervised settings. The PE algorithm has a computational advantage over conventional embedding methods based on pairwise object relations since its complexity scales with the product of the number of objects and the number of classes. We demonstrate PE by visualizing supervised categorization of Web pages, semisupervised categorization of digits, and the relations of words and latent topics found by an unsupervised algorithm, latent Dirichlet allocation.     

异质信息网络表征学习综述

随着信息技术的快速发展,信息网络无处不在,例如社交网络、学术网络、万维网等.由于网络规模不断扩大以及数据的稀疏性,信息网络的分析方法面临巨大挑战.作为应对网络规模及数据稀疏挑战的有效方法,信息网络表征学习旨在利用网络的拓扑结构、节点内容等信息将节点嵌入到低维的向量空间中,同时保留原始网络固有的结构特征和内容特征,从而使...     

半监督类保持局部线性嵌入方法

为使局部线性嵌入（local linear embedding, LLE）这一无监督高维数据的非线性特征提取方法提取出的特征在分类或聚类学习上更优,提出一种半监督类保持局部线性嵌入(semi-supervised class preserving local linear embedding, SSCLLE)的非线性特征提取方法。该方法将半监督信息融入到LLE中,首先对标记样本近邻赋予伪标签,增大标记样本数量。其次,对标记样本之间的距离进行局部调整,缩小同类样本间距,扩大异类样本间距。同时在局部线性嵌入优化目标函数中增加全局同类样本间距和异类样本间距的约束项,使得提取出的低维特征可以确保同类样本点互相靠近,而异类样本点彼此分离。在一系列实验中,其聚类精确度以及可视化效果明显高于无监督LLE和现有半监督流特征提取方法,表明该方法提取出的特征具有很好的类保持特性。     

特征注意金字塔调制网络的视频目标分割

目的 视频目标分割是在给定第1帧标注对象掩模条件下,实现对整个视频序列中感兴趣目标的分割。但是由于分割对象尺度的多样性,现有的视频目标分割算法缺乏有效的策略来融合不同尺度的特征信息。因此,本文提出一种特征注意金字塔调制网络模块用于视频目标分割。方法 首先利用视觉调制器网络和空间调制器网络学习分割对象的视觉和空间信息,并以此为先验引导分割模型适应特定对象的外观。然后通过特征注意金字塔模块挖掘全局上下文信息,解决分割对象多尺度的问题。结果 实验表明,在DAVIS 2016数据集上,本文方法在不使用在线微调的情况下,与使用在线微调的最先进方法相比,表现出更具竞争力的结果,J-mean指标达到了78.7%。在使用在线微调后,本文方法的性能在DAVIS 2017数据集上实现了最好的结果,J-mean指标达到了68.8%。结论 特征注意金字塔调制网络的视频目标分割算法在对感兴趣对象分割的同时,针对不同尺度的对象掩模能有效结合上下文信息,减少细节信息的丢失,实现高质量视频对象分割。     

基于联合知识表示学习的多模态实体对齐

基于知识表示学习的实体对齐方法是将多个知识图谱嵌入到低维语义空间,通过计算实体向量之间的相似度实现对齐.现有方法往往关注文本信息而忽视图像信息,导致图像中实体特征信息未得到有效利用.对此,提出一种基于联合知识表示学习的多模态实体对齐方法(ITMEA).该方法联合多模态(图像、文本)数据,采用TransE与TransD相结合的知识表示学习模型,使多模态数据能够嵌入到统一低维语义空间.在低维语义空间中迭代地学习已对齐多模态实体之间的关系,从而实现多模态数据的实体对齐.实验结果表明,ITMEA在WN18-IMG数据集中能够较好地实现多模态实体对齐.     

A fusion spatial attention approach for few-shot learning

Few-shot learning is a challenging problem in computer vision that aims to learn a new visual concept from very limited data. A core issue is that there is a large amount of uncertainty introduced by the small training set. For example, the few images may include cluttered backgrounds or different scales of objects. Existing approaches mostly address this problem from either the original image space or the embedding space by using meta-learning. To the best of our knowledge, none of them tackle this problem from both spaces jointly. To this end, we propose a fusion spatial attention approach that performs spatial attention in both image and embedding spaces. In the image space, we employ a Saliency Object Detection (SOD) module to extract the saliency map of an image and provide it to the network as an additional channel. In the embedding space, we propose an Adaptive Pooling (Ada-P) module tailored to few-shot learning that introduces a meta-learner to adaptively fuse local features of the feature maps for each individual embedding. The fusion process assigns different pooling weights to the features at different spatial locations. Then, weighted pooling can be conducted over an embedding to fuse local information, which can avoid losing useful information by considering the spatial importance of the features. The SOD and Ada-P modules can be used within a plug-and-play module and incorporated into various existing few-shot learning approaches. We empirically demonstrate that designing spatial attention methods for few-shot learning is a nontrivial task and our method has proven effective for it. We evaluate our method using both shallow and deeper networks on three widely used few-shot learning benchmarks, miniImageNet, tieredImageNet and CUB, and demonstrate very competitive performance.     

基于改进的等距离映射算法的人脸识别

针对等距离映射(Isomap)算法在处理扰动图像时拓扑结构不稳定的缺点,提出了一种改进算法。改进算法将图像欧氏距离(IMED)嵌入到等距离映射算法之中。首先引入坐标度量系数计算图像的坐标度量矩阵,通过线性变换将原始图像从欧氏距离(ED)空间转换到图像欧氏距离空间;然后计算变换空间中样本的欧氏距离矩阵,并在此基础上构建样本邻域图,得到近似测地距离矩阵;最后采用多维标度(MDS)分析算法构造样本的低维表示。对ORL和Yale人脸数据库降维并结合最近邻分类器进行实验,基于改进算法的识别率平均分别提高了5.57%和3.95%,表明与原算法相比,改进算法在人脸识别中对图像扰动具有较好的鲁棒性。     

基于选择关系嵌入算法的图像检索方法

结合流形学习和相关反馈技术的图像检索方法关键是结合低层可视化信息,从少量用户反馈信息中学习用户语义,以获得语义子空间流形。为获得更真实的语义子空间,文中在区分对待低层可视化和用户反馈信息的同时,基于低层可视化信息选择学习反馈信息中的类内和类间关系,提出一种选择关系嵌入算法应用于图像检索。该方法可保留更真实的语义流形结构,从而提高在低维空间中的检索精度。实验结果表明文中方法可将图像映射到更广范围的低维空间,在反馈迭代两次之后检索精度提高最高可达16。3%。     

Feature-Preserved 3D Canonical Form

Measuring the dissimilarity between non-rigid objects is a challenging problem in 3D shape retrieval. One potential solution is to construct the models’ 3D canonical forms (i.e., isometry-invariant representations in 3D Euclidean space) on which any rigid shape matching algorithm can be applied. However, existing methods, which are typically based on embedding procedures, result in greatly distorted canonical forms, and thus could not provide satisfactory performance to distinguish non-rigid models. In this paper, we present a feature-preserved canonical form for non-rigid 3D watertight meshes. The basic idea is to naturally deform original models against corresponding initial canonical forms calculated by Multidimensional Scaling (MDS). Specifically, objects are first segmented into near-rigid subparts, and then, through properly-designed rotations and translations, original subparts are transformed into poses that correspond well with their positions and directions on MDS canonical forms. Final results are obtained by solving nonlinear minimization problems for optimal alignments and smoothing boundaries between subparts. Experiments on two non-rigid 3D shape benchmarks not only clearly verify the advantages of our algorithm against existing approaches, but also demonstrate that, with the help of the proposed canonical form, we can obtain significantly better retrieval accuracy compared to the state of the art.     

Kernelized locality-sensitive hashing

Fast retrieval methods are critical for many large-scale and data-driven vision applications. Recent work has explored ways to embed high-dimensional features or complex distance functions into a low-dimensional Hamming space where items can be efficiently searched. However, existing methods do not apply for high-dimensional kernelized data when the underlying feature embedding for the kernel is unknown. We show how to generalize locality-sensitive hashing to accommodate arbitrary kernel functions, making it possible to preserve the algorithm's sublinear time similarity search guarantees for a wide class of useful similarity functions. Since a number of successful image-based kernels have unknown or incomputable embeddings, this is especially valuable for image retrieval tasks. We validate our technique on several data sets, and show that it enables accurate and fast performance for several vision problems, including example-based object classification, local feature matching, and content-based retrieval.     

结合时空一致性的FairMOT跟踪算法优化

目的 视频多目标跟踪(multiple object tracking, MOT)是计算机视觉中的一项重要任务,现有研究分别针对目标检测和目标关联部分进行改进,均忽视了多目标跟踪中的不一致问题。不一致问题主要包括3方面,即目标检测框中心与身份特征中心不一致、帧间目标响应不一致以及训练测试过程中相似度度量方式不一致。为了解决上述不一致问题,本文提出一种基于时空一致性的多目标跟踪方法,以提升跟踪的准确度。方法 从空间、时间以及特征维度对上述不一致性进行修正。对于目标检测框中心与身份特征中心不一致,针对每个目标检测框中心到特征中心之间的空间差异,在偏移后的位置上提取目标的ReID(re-identification)特征;对帧间响应不一致,使用空间相关计算相邻帧之间的运动偏移信息,基于该偏移信息对前一帧的目标响应进行变换后得到帧间一致性响应信息,然后对目标响应进行增强;对训练和测试过程中的相似度度量不一致,提出特征正交损失函数,在训练时考虑目标两两之间的相似关系。结果 在3个数据集上与现有方法进行比较。在MOT17、MOT20和Hieve数据集中,MOTA(multiple object t...     

边界流形嵌入在特征提取中的应用

样本点的边界信息对于分类具有重要意义。针对于边界Fisher分析（MFA）和局部敏感判别分析（LSDA）构造本征图和惩罚图所利用的样本点边界信息,在一些情况下并不能很好地表征不同类样本点的可分性,提出了一种新的图嵌入降维算法——边界流形嵌入（MME）。MME算法根据样本点的标签信息,寻找距离每个样本点最近的异类边界子流形,再返回本类中寻找距离异类边界子流形最近的同类边界子流形,从而定义出不同类样本间密切联系的同类边界邻域和异类边界邻域。通过最大化所有成对的边界子流形之间的距离,MME算法可以得到更具有鉴别意义的低维特征空间。同时,MME算法能将徘徊在边界的离群点收入到边界邻域里,这对减弱离群点给算法带来的负面的影响有一定的帮助。在人脸数据库上的实验结果表明了MME算法提取的低维特征能够提升分类的准确率。     

Grassmann multimodal implicit feature selection

In pattern recognition field, objects are usually represented by multiple features (multimodal features). For example, to characterize a natural scene image, it is essential to extract a set of visual features representing its color, texture, and shape information. However, integrating multimodal features for recognition is challenging because: (1) each feature has its specific statistical property and physical interpretation, (2) huge number of features may result in the curse of dimensionality (When data dimension is high, the distances between pairwise objects in the feature space become increasingly similar due to the central limit theory. This phenomenon influences negatively to the recognition performance), and (3) some features may be unavailable. To solve these problems, a new multimodal feature selection algorithm, termed Grassmann manifold feature selection (GMFS), is proposed. In particular, by defining a clustering criterion, the multimodal features are transformed into a matrix, and further treated as a point on the Grassmann manifold in Hamm and Lee (Grassmann discriminant analysis: a unifying view on subspace-based learning. In: Proceedings of the 25th international conference on machine learning (ICML), pp. 376–383, Helsinki, Finland [2008]). To deal with the unavailable features, L2-Hausdorff distance, a metric between different-sized matrices, is computed and the kernel is obtained accordingly. Based on the kernel, we propose supervised/unsupervised feature selection algorithms to achieve a physically meaningful embedding of the multimodal features. Experimental results on eight data sets validate the effectiveness the proposed approach.     

基于渐进式嵌套特征的融合网络

显著目标检测是指通过引入人类视觉注意力机制,使计算机能检测视觉场景中人们最感兴趣的区域或对象.针对显著性目标检测中存在检测边缘不清晰、检测目标不完整及小目标漏检的问题,文中提出基于渐进式嵌套特征的融合网络.网络采用渐进式压缩模块,将较深层特征不断向下传递融合,在降低模型参数量的同时也充分利用高级语义信息.先设计加权特征融合模块,将编码器的多尺度特征聚合成可访问高级信息和低级信息的特征图.再将聚合的特征分配到其它层,充分获取图像上下文信息及关注图像中的小目标对象.同时引入非对称卷积模块,进一步提高检测准确性.在6个公开数据集上的实验表明文中网络取得较优的检测效果.     

Mutual information-based method for selecting informative feature sets

Feature selection is one of the fundamental problems in pattern recognition and data mining. A popular and effective approach to feature selection is based on information theory, namely the mutual information of features and class variable. In this paper we compare eight different mutual information-based feature selection methods. Based on the analysis of the comparison results, we propose a new mutual information-based feature selection method. By taking into account both the class-dependent and class-independent correlation among features, the proposed method selects a less redundant and more informative set of features. The advantage of the proposed method over other methods is demonstrated by the results of experiments on UCI datasets (Asuncion and Newman, 2010 [1]) and object recognition.