Learning multi-task local metrics for image annotation

The goal of image annotation is to automatically assign a set of textual labels to an image to describe the visual contents thereof. Recently, with the rapid increase in the number of web images, nearest neighbor (NN) based methods have become more attractive and have shown exciting results for image annotation. One of the key challenges of these methods is to define an appropriate similarity measure between images for neighbor selection. Several distance metric learning (DML) algorithms derived from traditional image classification problems have been applied to annotation tasks. However, a fundamental limitation of applying DML to image annotation is that it learns a single global distance metric over the entire image collection and measures the distance between image pairs in the image-level. For multi-label annotation problems, it may be more reasonable to measure similarity of image pairs in the label-level. In this paper, we develop a novel label prediction scheme utilizing multiple label-specific local metrics for label-level similarity measure, and propose two different local metric learning methods in a multi-task learning (MTL) framework. Extensive experimental results on two challenging annotation datasets demonstrate that 1) utilizing multiple local distance metrics to learn label-level distances is superior to using a single global metric in label prediction, and 2) the proposed methods using the MTL framework to learn multiple local metrics simultaneously can model the commonalities of labels, thereby facilitating label prediction results to achieve state-of-the-art annotation performance.     

Overlap pattern synthesis with an efficient nearest neighbor classifier

Nearest neighbor (NN) classifier is the most popular non-parametric classifier. It is a simple classifier with no design phase and shows good performance. Important factors affecting the efficiency and performance of NN classifier are (i) memory required to store the training set, (ii) classification time required to search the nearest neighbor of a given test pattern, and (iii) due to the curse of dimensionality the number of training patterns needed by it to achieve a given classification accuracy becomes prohibitively large when the dimensionality of the data is high. In this paper, we propose novel techniques to improve the performance of NN classifier and at the same time to reduce its computational burden. These techniques are broadly based on: (i) overlap based pattern synthesis which can generate a larger number of artificial patterns than the number of input patterns and thus can reduce the curse of dimensionality effect, (ii) a compact representation of the given set of training patterns called overlap pattern graph (OLP-graph) which can be incrementally built by scanning the training set only once and (iii) an efficient NN classifier called OLP-NNC which directly works with OLP-graph and does implicit overlap based pattern synthesis. A comparison based on experimental results is given between some of the relevant classifiers. The proposed schemes are suitable for applications dealing with large and high dimensional datasets like those in data mining.     

基于自编码器的特征迁移算法

近年来,栈式自编码网络（stacked auto-encoder,SAE）在大规模数据集上表现出优异的图像分类性能。相对于其他图像分类方法中手工设计的低级特征,SAE的成功归因于深度网络能够学习到丰富的中级图像特征。然而,估计上百万个网络参数需要非常庞大的带标签的图像样本数据集。这样的性质阻止了SAE在小规模训练数据上的许多应用。在这篇文章中,提出的算法展示如何将SAE在大规模数据集上学习到的图像表示有效地迁移到只有有限训练数据的视觉识别任务中。实验部分设计了一个方法来复用在MNIST数据集上训练得到的隐藏层,以此计算在MNIST-variations数据集上的中级图像表示。实验结果展示了尽管两个数据集之间存在差异,但是被迁移的图像特征能够使得模型的分类性能得到极大的提升。     

A method of learning weighted similarity function to improve the performance of nearest neighbor

The performance of Nearest Neighbor (NN) classifier is known to be sensitive to the distance (or similarity) function used in classifying a test instance. Another major disadvantage of NN is that it uses all training instances in the generalization phase. This can cause slow execution speed and high storage requirement when dealing with large datasets. In the past research, many solutions have been proposed to handle one or both of the above problems. In the scheme proposed in this paper, we tackle both of these problems by assigning a weight to each training instance. The weight of a training instance is used in the generalization phase to calculate the distance (or similarity) of a query pattern to that instance. The basic NN classifier can be viewed as a special case of this scheme that treats all instances equally (by assigning equal weight to all training instances). Using this form of weighted similarity measure, we propose a learning algorithm that attempts to maximize the leave-one-out (LV1) classification rate of the NN rule by adjusting the weights of the training instances. At the same time, the algorithm reduces the size of the training set and can be viewed as a powerful instance reduction technique. An instance having zero weight is not used in the generalization phase and can be virtually removed from the training set. We show that our scheme has comparable or better performance than some recent methods proposed in the literature for the task of learning the distance function and/or prototype reduction.     

Representation learning with deep extreme learning machines for efficient image set classification

Efficient and accurate representation of a collection of images, that belong to the same class, is a major research challenge for practical image set classification. Existing methods either make prior assumptions about the data structure, or perform heavy computations to learn structure from the data itself. In this paper, we propose an efficient image set representation that does not make any prior assumptions about the structure of the underlying data. We learn the nonlinear structure of image sets with deep extreme learning machines that are very efficient and generalize well even on a limited number of training samples. Extensive experiments on a broad range of public datasets for image set classification show that the proposed algorithm consistently outperforms state-of-the-art image set classification methods both in terms of speed and accuracy.

     

融合扩充-双重特征提取应用于小样本学习

小样本图片分类的目标是根据极少数带有标注的样本去识别该类别,其中两个关键问题是带标注的数据量过少和不可见类别(训练类别和测试类别的不一致).针对这两个问题,我们提出了一个新的小样本分类模型:融合扩充-双重特征提取模型.首先,我们引入了一个融合扩充机制(FE),这个机制利用可见类别样本中同一类别不同样本之间的变化规则,对支持集的样本进行扩充,从而增加支持集中的样本数量,使提取的特征更具鲁棒性.其次,我们提出了一种双重特征提取机制(DF),该机制首先利用基类的大量数据训练两个不同的特征提取器:局部特征提取器和整体特征提取器,利用两个不同的特征提取器对样本特征进行提取,使提取的特征更加全面,然后根据局部和整体特征对比,突出对分类影响最大的特征,从而提高分类准确性.在Mini-ImageNet和Tiered-ImageNet数据集上,我们的模型都取得了较好的效果.     

基于加权共享近邻与累加序列的密度峰值算法

密度峰值聚类（DPC）算法在对密度分布差异较大的数据进行聚类时效果不佳,聚类结果受局部密度及其相对距离影响,且需要手动选取聚类中心,从而降低了算法的准确性与稳定性。为此,提出一种基于加权共享近邻与累加序列的密度峰值算法DPC-WSNN。基于加权共享近邻重新定义局部密度的计算方式,以避免截断距离选取不当对聚类效果的影响,同时有效处理不同类簇数据集分布不均的问题。在原有DPC算法决策值的基础上,生成一组累加序列,将累加序列的均值作为聚类中心和非聚类中心的临界点从而实现聚类中心的自动选取。利用人工合成数据集与UCI上的真实数据集测试与评估DPC-WSNN算法,并将其与FKNN-DPC、DPC、DBSCAN等算法进行比较,结果表明,DPC-WSNN算法具有更好的聚类表现,聚类准确率较高,鲁棒性较强。     

Combination of features through weighted ensembles for image classification

Image classification is a multi-class problem that is usually tackled with ensembles of binary classifiers. Furthermore, one of the most important challenges in this field is to find a set of highly discriminative image features for reaching a good performance in image classification. In this work we propose to use weighted ensembles as a method for feature combination. First, a set of binary classifiers are trained with a set of features and then, the scores are weighted with distances obtained from another set of feature vectors. We present two different approaches to weight the score vector: (1) directly multiplying each score by the weights and (2) fusing the scores values and the distances through a Neural Network. The experiments have shown that the proposed methodology improves classification accuracy of simple ensembles and even more it obtains similar classification accuracy than state-of-the-art methods, but using much less parameters.     

基于多代表点学习的RSKNN分类算法

RSKNN 算法是一种基于变精度粗糙集理论的 k-近邻改进算法,该算法能够保证在一定分类精度的前提下,有效地降低分类的计算量,提高分类效率。但由于 RSKNN 算法只是简单地将每个类中的样本划分成一个核心和边界区域,并没有根据数据集本身的特点进行划分,因而存在极大的局限性。针对存在的问题,提出一种多代表点学习算法,运用结构风险最小化理论对影响分类模型期望风险的因素进行分析,并使用无监督的局部聚类算法学习优化代表点集合。在UCI公共数据集上的实验表明,该算法比RSKNN算法具有更高的分类精度。     

IIRNet: A lightweight deep neural network using intensely inverted residuals for image recognition

Deep neural networks have achieved great success in many tasks of pattern recognition. However, large model size and high cost in computation limit their applications in resource-limited systems. In this paper, our focus is to design a lightweight and efficient convolutional neural network architecture by directly training the compact network for image recognition. To achieve a good balance among classification accuracy, model size, and computation complexity, we propose a lightweight convolutional neural network architecture named IIRNet for resource-limited systems. The new architecture is built based on Intensely Inverted Residual block (IIR block) to decrease the redundancy of the convolutional blocks. By utilizing two new operations, intensely inverted residual and multi-scale low-redundancy convolutions, IIR block greatly reduces its model size and computational costs while matches the classification accuracy of the state-of-the-art networks. Experiments on CIFAR-10, CIFAR-100, and ImageNet datasets demonstrate the superior performance of IIRNet on the trade-offs among classification accuracy, computation complexity, and model size, compared to the mainstream compact network architectures.     

提高小样本高光谱图像分类性能的变维卷积神经网络

目的 为了解决基于卷积神经网络的算法对高光谱图像小样本分类精度较低、模型结构复杂和计算量大的问题,提出了一种变维卷积神经网络。方法 变维卷积神经网络对高光谱分类过程可根据内部特征图维度的变化分为空—谱信息融合、降维、混合特征提取与空—谱联合分类的过程。这种变维结构通过改变特征映射的维度,简化了网络结构并减少了计算量,并通过对空—谱信息的充分提取提高了卷积神经网络对小样本高光谱图像分类的精度。结果 实验分为变维卷积神经网络的性能分析实验与分类性能对比实验,所用的数据集为Indian Pines和Pavia University Scene数据集。通过实验可知,变维卷积神经网络对高光谱小样本可取得较高的分类精度,在Indian Pines和Pavia University Scene数据集上的总体分类精度分别为87.87%和98.18%,与其他分类算法对比有较明显的性能优势。结论 实验结果表明,合理的参数优化可有效提高变维卷积神经网络的分类精度,这种变维模型可较大程度提高对高光谱图像中小样本数据的分类性能,并可进一步推广到其他与高光谱图像相关的深度学习分类模型中。     

跨数据集评估的高光谱图像分类

目的 随着高光谱成像技术的飞速发展,高光谱数据的应用越来越广泛,各场景高光谱图像的应用对高精度详细标注的需求也越来越旺盛。现有高光谱分类模型的发展大多集中于有监督学习,大多数方法都在单个高光谱数据立方中进行训练和评估。由于不同高光谱数据采集场景不同且地物类别不一致,已训练好的模型并不能直接迁移至新的数据集得到可靠标注,这也限制了高光谱图像分类模型的进一步发展。本文提出跨数据集对高光谱分类模型进行训练和评估的模式。方法 受零样本学习的启发,本文引入高光谱类别标签的语义信息,拟通过将不同数据集的原始数据及标签信息分别映射至同一特征空间以建立已知类别和未知类别的关联,再通过将训练数据集的两部分特征映射至统一的嵌入空间学习高光谱图像视觉特征和类别标签语义特征的对应关系,即可将该对应关系应用于测试数据集进行标签推理。结果 实验在一对同传感器采集的数据集上完成,比较分析了语义—视觉特征映射和视觉—语义特征映射方向,对比了5种基于零样本学习的特征映射方法,在高光谱图像分类任务中实现了对分类模型在不同数据集上的训练和评估。结论 实验结果表明,本文提出的基于零样本学习的高光谱分类模型可以实现跨数据集对分类模型进行训练和评估,在高光谱图像分类任务中具有一定的发展潜力。     

A co-boost framework for learning object categories from Google Images with 1st and 2nd order features

Conventional object recognition techniques rely heavily on manually annotated image datasets to achieve good performances. However, collecting high quality datasets is really laborious. The image search engines such as Google Images seem to provide quantities of object images. Unfortunately, a large portion of the search images are irrelevant. In this paper, we propose a semi-supervised framework for learning visual categories from Google Images. We exploit a co-training algorithm, the CoBoost algorithm, and integrate it with two kinds of features, the 1st and 2nd order features, which define bag of words representation and spatial relationship between local features, respectively. We create two boosting classifiers based on the 1st and 2nd order features in the training, during which one classifier provides labels for the other. The 2nd order features are generated dynamically rather than extracted exhaustively to avoid high computation. An active learning technique is also introduced to further improve the performance. Experimental results show that the object models learned from Google Images by our method are competitive with the state-of-the-art unsupervised approaches and some supervised techniques on the standard benchmark datasets.     

一种高精度高光谱图像分类方案设计

为了有效改善高光谱图像数据分类的精确度,减少对大数目数据集的依赖,在原型空间特征提取方法的基础上,提出一种基于加权模糊C均值算法改进型原型空间特征提取方案。该方案通过加权模糊C均值算法对每个特征施加不同的权重,从而保证提取后的特征含有较高的有效信息量,从而达到减少训练数据集而不降低分类所需信息量的效果。实验结果表明,与业内公认的原型空间提取算法相比,该方案在相对较小的数据集下,其性能仍具有较为理想的稳定性,且具有相对较高的分类精度。     

基于胶囊神经网络的车型精细识别研究

车辆型号精细识别在智能交通系统、涉车刑侦案件侦破等方面具有十分重要的应用前景。针对车辆型号种类繁多、部分型号区分度小等带来的车辆型号精细分类困难的问题,提出一种基于胶囊神经网络(capsule network,CapsNet)的车型图像识别模型CapCar。以CompCars数据集作为样本,首先通过加权平均值法进行图像的灰度化处理,减少数据集训练计算量,提高模型的训练速度。然后通过胶囊神经网络提取车型图像的全部特征和局部特征,实现车型分类识别。相较于现有的车型精细识别方法,该方法在提高识别精度的同时,有效压缩了模型参数规模。在基准数据集CompCars下进行大量实验的结果表明,CapCar模型进行车辆精细识别精度可达98.89%,其识别率高于一些其他经典的网络模型。CapCar模型参数大小仅为6.3 MB。该算法具有一定的先进性。     

联合结构相似性与类信息的图像分类

针对卷积神经网络训练收敛速度慢的问题，提出了一种加权的联合结构相似性和类信息监督训练的方法。首先，针对小图像，设计一个能有效提取图像高级别信息的卷积神经网络。其次，建立加权的联合结构相似性和类信息损失函数训练卷积神经网络。最后，通过mnist手写数字和cifar10图像分类实验验证所设计网络的有效性。实验结果表明，所设计的网络在mnist手写数字和cifar10数据集上的图像分类错误率分别为0.33%和11%。在未进行扩增mnist数据集的前提下，所设计的网络的性能超过了该数据集上所有单网络的性能；在cifar10数据集上，所设计的网络能以较少的计算量获得较高的图像分类准确率。同时，联合结构相似性和类信息损失的监督训练能加快网络的训练速度。     

基于小样本学习的SAR图像识别

深度学习已成为图像识别领域的一个研究热点。与传统图像识别方法不同,深度学习从大量数据中自动学习特征,并且具有强大的自学习能力和高效的特征表达能力。但在小样本条件下,传统的深度学习方法如卷积神经网络难以学习到有效的特征,造成图像识别的准确率较低。因此,提出一种新的小样本条件下的图像识别算法用于解决SAR图像的分类识别。该算法以卷积神经网络为基础,结合自编码器,形成深度卷积自编码网络结构。首先对图像进行预处理,使用2D Gabor滤波增强图像,在此基础上对模型进行训练,最后构建图像分类模型。该算法设计的网络结构能自动学习并提取小样本图像中的有效特征,进而提高识别准确率。在MSTAR数据集的10类目标分类中,选择训练集数据中10%的样本作为新的训练数据,其余数据为验证数据,并且,测试数据在卷积神经网络中的识别准确率为76.38%,而在提出的卷积自编码结构中的识别准确率达到了88.09%。实验结果表明,提出的算法在小样本图像识别中比卷积神经网络模型更加有效。     

基于语义标签生成和偏序结构的图像层级分类

智能电子设备和互联网的普及,使得图像数据爆炸性膨胀.为了有效管理复杂图像资源,本文提出了一种基于加权语义邻近集和形式概念偏序结构的图像层级分类方法.首先,根据图像语义相关分数,对不同程度语义设定自适应权系数,从训练图库中构建加权语义邻近集,通过对语义邻近集中图像的词频分布进行判决,自动生成图像的多个语义标签;然后,以每幅图像为对象,以每幅图像自动生成的语义标签为属性,构建形式背景,通过偏序结构算法对复杂图像集进行有效的层级分类.本文方法可以得到图像库中图像之间明确的结构关系和图像类别之间的从属关系,为复杂图像大数据进行层级分类管理提供了有效的思路.本文对Corel5k、EspGame和Iaprtc12三个数据库进行了图像标注实验,证明了标注的语义完整性和主要语义的准确性;并对Corel5k数据库进行了图像的层级分类实验,结果表明层级分类效果显著.