Clustering using PK-D: A connectivity and density dissimilarity

We present a new dissimilarity, which combines connectivity and density information. Usually, connectivity and density are conceived as mutually exclusive concepts; however, we discuss a novel procedure to merge both information sources. Once we have calculated the new dissimilarity, we apply MDS in order to find a low dimensional vector space representation. The new data representation can be used for clustering and data visualization, which is not pursued in this paper. Instead we use clustering to estimate the gain from our approach consisting of dissimilarity + MDS. Hence, we analyze the partitions’ quality obtained by clustering high dimensional data with various well known clustering algorithms based on density, connectivity and message passing, as well as simple algorithms like k-means and Hierarchical Clustering (HC). The quality gap between the partitions found by k-means and HC alone compared to k-means and HC using our new low dimensional vector space representation is remarkable. Moreover, our tests using high dimensional gene expression and image data confirm these results and show a steady performance, which surpasses spectral clustering and other algorithms relevant to our work.     

基于颜色的压缩层次图像表示方法

空间金字塔模型在每层中把图像划分成细胞单元用于给图像表示提供空间信息,但是这种方式不能很好地匹配对象上的不同部分,为此提出一种基于颜色的层次（CL）划分算法。CL算法从多特征融合的角度出发,通过优化的方式在不同层次中得到每个类别中有判别力的颜色,然后根据每层中有判别力的颜色对图像进行迭代的层次划分;最后连接不同层次直方图用于图像表示。为了解决图像表示维度过高的问题,采用分化信息理论的特征聚类（DITC）方法对字典进行聚类用于字典降维,并用压缩生成的字典进行最终的图像表示。实验结果表明,所提方法能够在Soccer,Flower 17和Flower 102上取得良好的识别效果。     

遥感图像变化区域的无监督压缩感知

传统的基于结构特征的遥感图像变化检测方法,易受成像稳定性的影响而误差很大。针对图像内在的稀疏性结构信息,提出基于压缩感知(CS)的遥感图像变化检测方法。通过自适应构造超完备字典将图像局部信息投影到高维空间中,实现图像的稀疏表示,并运用随机矩阵得到了数据在高维空间中的低维特征子空间。最后利用模糊C均值(FCM)聚类算法进行无监督聚类,实现遥感图像变化区域信息的重构。实验结果表明,本文方法不仅能够很好的检测出图像的轮廓变化和图像的区域变化,而且对噪声具有很好的鲁棒性。     

Building global image features for scene recognition

In this paper, we present a simple, yet very efficient global image representation for scene recognition. A scene image is represented by a histogram of local transforms, which is an extended version of census transform histogram. The local transforms include local difference sign and magnitude information. Due to strong constraints between neighboring transformed values, global structure information can be captured through the histogram and spatial pyramid representation. Principal component analysis is used to reduce the dimensionality and get a compact feature vector. Experimental results on three widely used datasets demonstrate that the proposed method could achieve competitive performance in terms of speed and accuracy.     

基于多尺度特征映射匹配的图像表示方法

在卷积神经网络模型中,空间金字塔池化方法将空间信息融入到深度特征的生成过程中,最终生成的图像表示可以有效地用于提高图像检索性能,但是此方法会导致生成的图像表示中不同维度之间描述的信息存在重复且相同维度描述的图像内容不匹配。为此提出了一种基于多尺度特征映射匹配（multi-scale feature map matching,MFMM）的图像表示方法,此方法首先利用深度特征的方差与协方差矩阵提出了一种特征映射选择算法,用于增强图像表示中不同维度特征的独立性。其次,依据相同通道特征映射中高响应值位置有较高匹配性的特点,结合激活映射中最大响应位置的深度特征提出了一种优化的特征映射中心点选择方法。最后,按照不同的中心点通过多尺度窗口采样的方式,从特征映射中提取出带有空间信息的深度特征用于表示图像内容。实验结果表明,提出的方法在图像检索任务中能够取得良好的效果。     

基于低秩稀疏评分的非监督特征选择

在处理高维数据过程中,特征选择是一个非常重要的数据降维步骤。低秩表示模型具有揭示数据全局结构信息的能力和一定的鉴别能力。稀疏表示模型能够利用较少的连接关系揭示数据的本质结构信息。在低秩表示模型的基础上引入稀疏约束项,构建一种低秩稀疏表示模型学习数据间的低秩稀疏相似度矩阵;基于该矩阵提出一种低秩稀疏评分机制用于非监督特征选择。在不同数据库上将选择后的特征进行聚类和分类实验,同传统特征选择算法进行比较。实验结果表明了低秩特征选择算法的有效性。     

Multiresolution-based watersheds for efficient image segmentation

This paper presents an efficient method for image segmentation based on a multiresolution application of a wavelet transform and watershed segmentation algorithm. The procedure toward complete segmentation consists of four steps: pyramid representation, image segmentation, region merging and region projection. First, pyramid representation creates multiresolution images using a wavelet transform. Second, image segmentation segments the lowest-resolution image of the pyramid using a watershed segmentation algorithm. Third, region merging merges the segmented regions using the third-order moment values of the wavelet coefficients. Finally, the segmented low-resolution image with label is projected into a full-resolution image (original image) by inverse wavelet transform. Experimental results of the presented method can be applied to the segmentation of noise or degraded images as well as reduce over-segmentation. In addition, we applied our method to human face detection with accurate and closed boundaries.     

基于多视觉码本的图像表示

基于词袋模型的图像表示方法的有效性主要受限于局部特征的量化误差。文中提出一种基于多视觉码本的图像表示方法,通过综合考虑码本构建和编码方法这两个方面的因素加以改进。具体包括:1)多视觉码本构建,以迭代方式构建多个紧凑且具有互补性的视觉码本;2)图像表示,首先针对多码本的情况,依次从各码本中选择相应的视觉单词并采用线性回归估计编码系数,然后结合图像的空间金字塔结构形成最终的图像表示。在一些标准测试集合的图像分类结果验证文中方法的有效性。     

基于语义特征空间上下文的短文本表示学习

文本表示是自然语言处理中的基础任务,针对传统短文本表示高维稀疏问题,提出1种基于语义特征空间上下文的短文本表示学习方法。考虑到初始特征空间维度过高,通过计算词项间互信息与共现关系,得到初始相似度并对词项进行聚类,利用聚类中心表示降维后的语义特征空间。然后,在聚类后形成的簇上结合词项的上下文信息,设计3种相似度计算方法分别计算待表示文本中词项与特征空间中特征词的相似度,以形成文本映射矩阵对短文本进行表示学习。实验结果表明,所提出的方法能很好地反映短文本的语义信息,能对短文本进行合理而有效的表示学习。     

高显著性的时空金字塔精简描述符算法研究

当传统时空金字塔层数较多时,特征描述符的维数会非常高,使得此类描述符在训练和测试阶段计算效率非常低。此外,在时空金字塔的分层及每层立方体单元的划分中,至今仍然采用手动划分视频,使得视频划分策略没有强的理论依据。鉴于以上缺点,提出一种高显著性的时空金字塔精简描述符算法。形成的新描述符是所有时空金字塔层中每个立方体单元局部特征的权重和,而不是把所有立方体单元局部特征描述符串联起来形成一个巨大的特征描述符,每个立方体单元的权重可以通过偏最小二乘法自动获取,由此产生的视频全局描述符精简并且具有高的显著性。此外通过观测其精细立方体单元的权重,还可以展现出显著性时空金字塔每个立方体单元及每层金字塔的贡献,由此,可以根据权重自动划分视频。采用HMDB51和YouTube两个动作数据库进行实验验证,与时空金字塔描述符和超稀疏编码向量相比,此描述符精简并能在低维度下取得较好的识别效果。     

基于多维金字塔表达和AdaBoost的高分辨率SAR图像城区场景分类算法

提出了多维金字塔表达算法, 并使用基于多维金字塔表达的AdaBoost实现了高分辨率合成孔径雷达(Synthetic aperture radar, SAR)图像的城区场景分类. 多维金字塔表达算法首先在局部特征的各维计算金字塔表达矢量, 再将所有的金字塔表达矢量连接起来构成多维金字塔表达矢量. 多维金字塔表达算法克服了金字塔表达算法在处理高维局部特征时, 遇到的输出金字塔表达矢量的区分力受计算效率制约的问题. 本文分别在一个TerraSAR-X图像库和一张大幅TerraSAR-X图像上比较基于金字塔表达的AdaBoost和基于多维金字塔表达的AdaBoost的分类性能. 实验结果表明, 与前者相比, 后者显著提高了计算效率同时保证了分类精度.     

Use of Spectral Clustering Combined with Normalized Cuts (N-Cuts) in an Iterative k-Means Clustering Framework (NKSC) for Superpixel Segmentation with Contour Adherence

Superpixel segmentation methods are generally used as a pre-processing step to speed up image processing tasks. They group the pixels of an image into homogeneous regions while trying to respect existing contours. In this paper, we propose a fast Superpixels segmentation algorithm with Contour Adherence using spectral clustering, combined with normalized cuts in an iterative k-means clustering framework. It produces compact and uniform superpixels with low computational costs. Normalized cut is adapted to measure the color similarity and space proximity between image pixels. We have used a kernel function to estimate the similarity metric. Kernel function maps the pixel values and coordinates into a high dimensional feature space. The objective functions of weighted K-means and normalized cuts share the same optimum point in this feature space. So it is possible to optimize the cost function of normalized cuts by iteratively applying simple K-means clustering algorithm. The proposed framework produces regular and compact superpixels that adhere to the image contours. On segmentation comparison benchmarks it proves to be equally well or better than the state-of-the-art super pixel segmentation algorithms in terms of several commonly used evaluation metrics in image segmentation. In addition, our method is computationally very efficient and its computational complexity is linear.     

Different Aspects of Clustering The Self-Organizing Maps

Self-organizing map (SOM) is an artificial neural network tool that is trained using unsupervised learning to produce a low dimensional representation of the input space, called a map. This map is generally the object of a clustering analysis step which aims to partition the referents vectors (map neurons) into compact and well-separated groups. In this paper, we consider the problem of the clustering SOM using different aspects: partitioning, hierarchical and graph coloring based techniques. Unlike the traditional clustering SOM techniques, which use k-means or hierarchical clustering, the graph-based approaches have the advantage of providing a partitioning of the SOM by simultaneously using dissimilarities and neighborhood relations provided by the map. We present the experimental results of several comparisons between these different ways of clustering.     

基于模糊聚类规则的图像去雾方法研究

有雾图像的复原问题已经成为了机器视觉领域研究的热点。在传统图像复原理论的基础上,提出一种新的基于模糊聚类规则图像去雾方法,用于恢复被雾影响的降质图像。利用模糊聚类的方法对有雾图像进行预处理,从而获取图像复原过程中的透射率参数。为获取有效的景深信息,将透射率图像与提取到的图像饱和度效果图通过多尺度金字塔细化策略进行有效融合,从而获得图像景深信息。最后,通过提取到的透射率和景深信息实现图像的复原。对比实验效果显示,提出的图像去雾算法能够有效去除图像中的雾信息,同时算法的计算速度相对较快。     

A scalable algorithm for extraction and clustering of event-related pictures

The event detection problem, which is closely related to clustering, has gained a lot of attentions within event detection for textual documents. However, although image clustering is a problem that has been treated extensively in both Content-Based Image Retrieval (CBIR) and Text-Based Image Retrieval (TBIR) systems, event detection within image management is a relatively new area. Having this in mind, we propose a novel approach for event extraction and clustering of images, taking into account textual annotations, time and geographical positions. Our goal is to develop a clustering method based on the fact that an image may belong to an event cluster. Here, we stress the necessity of having an event clustering and cluster extraction algorithm that are both scalable and allow online applications. To achieve this, we extend a well-known clustering algorithm called Suffix Tree Clustering (STC), originally developed to cluster text documents using document snippets. The idea is that we consider an image along with its annotation as a document. Further, we extend it to also include time and geographical position so that we can capture the contextual information from each image during the clustering process. This has appeared to be particularly useful on images gathered from online photo-sharing applications such as Flickr. Hence, our STC-based approach is aimed at dealing with the challenges induced by capturing contextual information from Flickr images and extracting related events. We evaluate our algorithm using different annotated datasets mainly gathered from Flickr. As part of this evaluation we investigate the effects of using different parameters, such as time and space granularities, and compare these effects. In addition, we evaluate the performance of our algorithm with respect to mining events from image collections. Our experimental results clearly demonstrate the effectiveness of our STC-based algorithm in extracting and clustering events.     

非负子空间聚类指导下的非负矩阵分解

非负矩阵分解作为一种有效的数据表示方法被广泛应用于模式识别和机器学习领域。为了得到原始数据紧致有效的低维数据表示,无监督非负矩阵分解方法在特征降维的过程中通常需要同时发掘数据内部隐含的几何结构信息。通过合理建模数据样本间的相似性关系而构建的相似度图,通常被用来捕获数据样本的空间分布结构信息。子空间聚类可以有效发掘数据内部的子空间结构信息,其获得的自表达系数矩阵可用于构建相似度图。该文提出了一种非负子空间聚类算法来发掘数据的子空间结构信息,同时利用该信息指导非负矩阵分解,从而得到原始数据有效的非负低维表示。同时,该文还提出了一种有效的迭代求解方法来求解非负子空间聚类问题。在两个图像数据集上的聚类实验结果表明,利用数据的子空间结构信息可以有效改善非负矩阵分解的性能。     

Image representation using Laplacian regularized nonnegative tensor factorization

Tensor provides a better representation for image space by avoiding information loss in vectorization. Nonnegative tensor factorization (NTF), whose objective is to express an n-way tensor as a sum of k rank-1 tensors under nonnegative constraints, has recently attracted a lot of attentions for its efficient and meaningful representation. However, NTF only sees Euclidean structures in data space and is not optimized for image representation as image space is believed to be a sub-manifold embedded in high-dimensional ambient space. To avoid the limitation of NTF, we propose a novel Laplacian regularized nonnegative tensor factorization (LRNTF) method for image representation and clustering in this paper. In LRNTF, the image space is represented as a 3-way tensor and we explicitly consider the manifold structure of the image space in factorization. That is, two data points that are close to each other in the intrinsic geometry of image space shall also be close to each other under the factorized basis. To evaluate the performance of LRNTF in image representation and clustering, we compare our algorithm with NMF, NTF, NCut and GNMF methods on three standard image databases. Experimental results demonstrate that LRNTF achieves better image clustering performance, while being more insensitive to noise.     

基于多尺度LBP金字塔特征的分类算法

为有效解决旋转变化、光照变化和尺度变化等图像的分类问题,提出一种基于多尺度局部二元模式(LBP)金字塔特征的图像分类算法。通过多尺度LBP金字塔提取各尺度的图像纹理特征,建立图像的多尺度LBP金字塔直方图,并将其作为图像特征向量,采用K-means方法对该特征向量进行降维,以用于图像分类。同时,针对传统二进制权值分布方法对噪声敏感的缺点,提出一种多端权值分布方法。实验结果表明,多尺度LBP金字塔方法具有较好的可鉴别性及图像描述能力,而多端权值分布法也能提高图像的分类精度。     

图正则化非负矩阵分解的异质网社区发现

挖掘数据网络中有价值的、具有稳定性的社区,对网络信息的获取、推荐及网络的演化预测具有重要的价值。针对现有异质网络聚类方法难以在同一维度有效整合网络中异质信息的问题,提出了一种基于图正则化非负矩阵分解的异质网络聚类方法。通过加入图正则项,将中心类型子空间和属性类型子空间的内部连接关系作为约束项,引入到非负矩阵分解模型中,从而找到高维数据在低维空间的紧致嵌入,成功消除了异质节点之间的部分噪声,同时,对反映不同子网络共有潜在结构的共识矩阵进行优化,有效整合异质信息,并且在降维过程中较大限度地保留了异质信息的完整性,提高了异质网络聚类方法的精度,在真实世界数据集上的实验结果也验证了该方法的有效性。