Two-Tier genetic programming: towards raw pixel-based image classification

Classifying images is of great importance in machine vision and image analysis applications such as object recognition and face detection. Conventional methods build classifiers based on certain types of image features instead of raw pixels because the dimensionality of raw inputs is often too large. Determining an optimal set of features for a particular task is usually the focus of conventional image classification methods. In this study we propose a Genetic Programming (GP) method by which raw images can be directly fed as the classification inputs. It is named as Two-Tier GP as every classifier evolved by it has two tiers, the other for computing features based on raw pixel input, one for making decisions. Relevant features are expected to be self-constructed by GP along the evolutionary process. This method is compared with feature based image classification by GP and another GP method which also aims to automatically extract image features. Four different classification tasks are used in the comparison, and the results show that the highest accuracies are achieved by Two-Tier GP. Further analysis on the evolved solutions reveals that there are genuine features formulated by the evolved solutions which can classify target images accurately.     

Image mining by spectral features: A case study of scenery image classification

Spectral features of images, such as Gabor filters and wavelet transform can be used for texture image classification. That is, a classifier is trained based on some labeled texture features as the training set to classify unlabeled texture features of images into some pre-defined classes. The aim of this paper is twofold. First, it investigates the classification performance of using Gabor filters, wavelet transform, and their combination respectively, as the texture feature representation of scenery images (such as mountain, castle, etc.). A k-nearest neighbor (k-NN) classifier and support vector machine (SVM) are also compared. Second, three k-NN classifiers and three SVMs are combined respectively, in which each of the combined three classifiers uses one of the above three texture feature representations respectively, to see whether combining multiple classifiers can outperform the single classifier in terms of scenery image classification. The result shows that a single SVM using Gabor filters provides the highest classification accuracy than the other two spectral features and the combined three k-NN classifiers and three SVMs.     

A survey of document image classification: problem statement,classifier architecture and performance evaluation

Document image classification is an important step in Office Automation, Digital Libraries, and other document image analysis applications. There is great diversity in document image classifiers: they differ in the problems they solve, in the use of training data to construct class models, and in the choice of document features and classification algorithms. We survey this diverse literature using three components: the problem statement, the classifier architecture, and performance evaluation. This brings to light important issues in designing a document classifier, including the definition of document classes, the choice of document features and feature representation, and the choice of classification algorithm and learning mechanism. We emphasize techniques that classify single-page typeset document images without using OCR results. Developing a general, adaptable, high-performance classifier is challenging due to the great variety of documents, the diverse criteria used to define document classes, and the ambiguity that arises due to ill-defined or fuzzy document classes.     

应用级联多分类器的高光谱图像分类

目的 高光谱分类任务中,由于波段数量较多,图像中存在包含噪声以及各类地物样本分布不均匀等问题,导致分类精度与训练效率不能平衡,在小样本上分类精度低。因此,提出一种基于级联多分类器的高光谱图像分类方法。方法 首先采用主成分分析方法将高度相关的高维特征合成无关的低维特征,以加快Gabor滤波器提取纹理特征的速度;然后使用Gabor滤波器提取图像在各个尺寸、方向上的纹理信息,每一个滤波器会生成一张特征图,在特征图中以待分类样本为中心取一个d×d的邻域,计算该邻域内数据的均值和方差来作为待分类样本的空间信息,再将空间信息和光谱信息融合,以降低光线与噪声的影响;最后将谱—空联合特征输入级联多分类器中,得到预测样本关于类别的概率分布的平均值。结果 实验采用Indian Pines、Pavia University和Salinas 3个数据集,与经典算法如支持向量机和卷积神经网络进行比较,并利用总体分类精度、平均分类精度和Kappa系数作为评价标准进行分析。本文方法总体分类精度在3个数据集上分别达到97.24%、99.57%和99.46%,相对于基于径向基神经网络（RBF）核函数的支持向量机方法提高了13.2%、4.8%和5.68%,相对于加入谱—空联合特征的RBF-SVM （radial basis function-support vector machine）方法提高了2.18%、0.36%和0.83%,相对于卷积神经网络方法提高了3.27%、3.2%和0.3%;Kappa系数分别是0.968 6、0.994 3和0.995 6,亦有提高。结论 实验结果表明,本文方法应用于高光谱图像分类具有较优的分类效果,训练效率较高,无需依赖GPU,而且在小样本上也具有较高的分类精度。     

Hyperspectral image classification using multiobjective optimization

Hyperspectral images constitute a substantial amount of data in the form of spectral bands. This information is used for land cover analysis, specifically in classifying a hyperspectral pixel, which is a popular domain in remote sensing. This paper proposed an efficient framework to classify spectral-spatial hyperspectral images by employing multiobjective optimization. Spectral-spatial features of hyperspectral images are passed for optimization. As hyperspectral images have a high dimensional feature set, many classifiers cannot perform well. Multiobjective optimization reduces the feature set without affecting the discrimination ability of the classifier. The proposed work is validated on a standard hyperspectral image set, Pavia University and Kennedy Space Centre.

     

Fuzzy-rough feature selection aided support vector machines for Mars image classification

This paper presents a novel application of advanced machine learning techniques for Mars terrain image classification. Fuzzy-rough feature selection (FRFS) is adapted and then employed in conjunction with Support Vector Machines (SVMs) to construct image classifiers. These techniques are integrated to address problems in space engineering where the images are of many classes, large-scale, and diverse representational properties. The use of the adapted FRFS allows the induction of low-dimensionality feature sets from feature patterns of a much higher dimensionality. To evaluate the proposed work, K-Nearest Neighbours (KNNs) and decision trees (DTREEs) based image classifiers as well as information gain rank (IGR) based feature selection are also investigated here, as possible alternatives to the underlying machine learning techniques adopted. The results of systematic comparative studies demonstrate that in general, feature selection improves the performance of classifiers that are intended for use in high dimensional domains. In particular, the proposed approach helps to increase the classification accuracy, while enhancing classification efficiency by requiring considerably less features. This is evident in that the resultant SVM-based classifiers which utilise FRFS-selected features generally outperform KNN and DTREE based classifiers and those which use IGR-returned features. The work is therefore shown to be of great potential for on-board or ground-based image classification in future Mars rover missions.     

Handwritten digit recognition: benchmarking of state-of-the-art techniques

This paper presents the results of handwritten digit recognition on well-known image databases using state-of-the-art feature extraction and classification techniques. The tested databases are CENPARMI, CEDAR, and MNIST. On the test data set of each database, 80 recognition accuracies are given by combining eight classifiers with ten feature vectors. The features include chaincode feature, gradient feature, profile structure feature, and peripheral direction contributivity. The gradient feature is extracted from either binary image or gray-scale image. The classifiers include the k-nearest neighbor classifier, three neural classifiers, a learning vector quantization classifier, a discriminative learning quadratic discriminant function (DLQDF) classifier, and two support vector classifiers (SVCs). All the classifiers and feature vectors give high recognition accuracies. Relatively, the chaincode feature and the gradient feature show advantage over other features, and the profile structure feature shows efficiency as a complementary feature. The SVC with RBF kernel (SVC-rbf) gives the highest accuracy in most cases but is extremely expensive in storage and computation. Among the non-SV classifiers, the polynomial classifier and DLQDF give the highest accuracies. The results of non-SV classifiers are competitive to the best ones previously reported on the same databases.     

Detecting siblings in image pairs

In everyday life, face similarity is an important kinship clue. Computer algorithms able to infer kinship from pairs of face images could be applied in forensics, image retrieval and annotation, and historical studies. So far, little work in this area has been presented, and only one study, using a small set of low quality images, tackles the problem of identifying siblings pairs. The purpose of our paper is to present a comprehensive investigation on this subject, aimed at understanding which are, on the average, the most relevant facial features, how effective can be computer algorithms for detecting siblings pairs, and if they can outperform human evaluation. To avoid problems due to low quality pictures and uncontrolled imaging conditions, as for the heterogeneous datasets collected for previous researches, we prepared a database of high quality pictures of sibling pairs, shot in controlled conditions and including frontal, profile, expressionless, and smiling faces. Then we constructed various classifiers of image pairs using different types of facial data, based on various geometric, textural, and holistic features. The classifiers were first tested separately, and then the most significant facial data, selected with a two stage feature selection algorithm were combined into a unique classifier. The discriminating ability of the automatic classifier combining features of different nature has been found to outperform that of a panel of human raters. We also show the good generalization capabilities of the algorithm by applying the classifier, in a cross-database experiment, to a low quality database of images collected from the Internet.     

An Automated Brain Image Analysis System for Brain Cancer using Shearlets

In this paper, an Automated Brain Image Analysis (ABIA) system that classifies the Magnetic Resonance Imaging (MRI) of human brain is presented. The classification of MRI images into normal or low grade or high grade plays a vital role for the early diagnosis. The Non-Subsampled Shearlet Transform (NSST) that captures more visual information than conventional wavelet transforms is employed for feature extraction. As the feature space of NSST is very high, a statistical t-test is applied to select the dominant directional sub-bands at each level of NSST decomposition based on sub-band energies. A combination of features that includes Gray Level Co-occurrence Matrix (GLCM) based features, Histograms of Positive Shearlet Coefficients (HPSC), and Histograms of Negative Shearlet Coefficients (HNSC) are estimated. The combined feature set is utilized in the classification phase where a hybrid approach is designed with three classifiers; k-Nearest Neighbor (kNN), Naive Bayes (NB) and Support Vector Machine (SVM) classifiers. The output of individual trained classifiers for a testing input is hybridized to take a final decision. The quantitative results of ABIA system on Repository of Molecular Brain Neoplasia Data (REMBRANDT) database show the overall improved performance in comparison with a single classifier model with accuracy of 99% for normal/abnormal classification and 98% for low and high risk classification.     

基于模式分类的图像区域作伪检测

为了鉴别一幅数字图像是否存在作伪的区域,提出一种利用改进的图像特征进行区域作伪检测的算法.基于模式分类的思想,该方法把图像分割成适当大小的块,从图像块中提取特征数据,用SVM分类器训练数据并得到支持向量机模型,利用该模型检测嫌疑图片是否存在作伪.该算法从噪声相关性、残差噪声、图像质量、小波域等方面分析相机图片的特点,获取每种的统计特征,形成特征集.实验结果表明,该方法能有效地检测出图像的具体作伪区域.     

3D卷积自编码器高光谱图像分类模型

目的 高光谱图像分类是遥感领域的基础问题，高光谱图像同时包含丰富的光谱信息和空间信息，传统模型难以充分利用两种信息之间的关联性，而以卷积神经网络为主的有监督深度学习模型需要大量标注数据，但标注数据难度大且成本高。针对现有模型的不足，本文提出了一种无监督范式下的高光谱图像空谱融合方法，建立了3D卷积自编码器（3D convolutional auto-encoder，3D-CAE）高光谱图像分类模型。方法 3D卷积自编码器由编码器、解码器和分类器构成。将高光谱数据预处理后，输入到编码器中进行无监督特征提取，得到一组特征图。编码器的网络结构为3个卷积块构成的3D卷积神经网络，卷积块中加入批归一化技术防止过拟合。解码器为逆向的编码器，将提取到的特征图重构为原始数据，用均方误差函数作为损失函数判断重构误差并使用Adam算法进行参数优化。分类器由3层全连接层组成，用于判别编码器提取到的特征。以3D-CNN （three dimensional convolutional neural network）为自编码器的主干网络可以充分利用高光谱图像的空间信息和光谱信息，做到空谱融合。以端到端的方式对模型进行训练可以省去复杂的特征工程和数据预处理，模型的鲁棒性和稳定性更强。结果 在Indian Pines、Salinas、Pavia University和Botswana等4个数据集上与7种传统单特征方法及深度学习方法进行了比较，本文方法均取得最优结果，总体分类精度分别为0.948 7、0.986 6、0.986 2和0.964 9。对比实验结果表明了空谱融合和无监督学习对于高光谱遥感图像分类的有效性。结论 本文模型充分利用了高光谱图像的光谱特征和空间特征，可以做到无监督特征提取，无需大量标注数据的同时分类精度高，是一种有效的高光谱图像分类方法。     

分类与分割相结合的JPEG图像隐写分析

目的 隐写分析研究现状表明,与秘密信息的嵌入过程相比,图像内容和统计特性差异对隐写检测特征分布会造成更大的影响,这导致图像隐写分析成为了一个"相同类内特征分布分散、不同类间特征混淆严重"的分类问题。针对此问题,提出了一种更加有效的JPEG图像隐写检测模型。方法 通过对隐写检测常用的分类器进行分析,从降低隐写检测特征类内离散度的角度入手,将基于图像内容复杂度的预分类和图像分割相结合,根据图像内容复杂度对图像进行分类、分割,然后分别对每一类子图像提取高维富模型隐写检测特征,构建分类器进行训练和测试,并通过加权融合得到最终的检测结果。结果 在实验部分,对具有代表性的隐写检测特征集提取了两类可分性判据,对本文算法的各类别、区域所提取特征的可分性均得到明显提高,证明了模型的有效性。同时在训练、测试图像库匹配和不匹配的情况下,对算法进行了二分类测试,并与其他算法进行了性能比较,本文算法的检测性能均有所提高,性能提升最高接近10%。结论 本文算法能够有效提高隐写检测性能,尤其是在训练、测试图像库统计特性不匹配的情况下,本文算法性能提升更加明显,更适合于实际复杂网络下的应用。     

Supervised image segmentation using Q-Shift Dual-Tree Complex Wavelet Transform coefficients with a texton approach

In this study, we propose a simple and efficient texture-based algorithm for image segmentation. This method constitutes computing textons and bag of words (BOWs) learned by support vector machine (SVM) classifiers. Textons are composed of local magnitude coefficients that arise from the Q-Shift Dual-Tree Complex Wavelet Transform (DT-CWT) combined with color components. In keeping with the needs of our research context, which addresses land cover mapping from remote images, we use a few small texture patches at the training stage, where other supervised methods usually train fully representative textures. We accounted for the scale and rotation invariance issue of the textons, and three different invariance transforms were evaluated on DT-CWT-based features. The largest contribution of this study is the comparison of three classification schemes in the segmentation algorithm. Specifically, we designed a new scheme that was especially competitive and that uses several classifiers, with each classifier adapted to a specific size of analysis window in texton quantification and trained on a reduced data set by random selection. This configuration allows quick SVM convergence and an easy parallelization of the SVM-bank while maintaining a high segmentation accuracy. We compare classification results with textons made using the well-known maximum response filters bank and speed up robust features features as references. We show that DT-CWT textons provide better distinguishing features in the entire set of configurations tested. Benchmarks of our different method configurations were made over two substantial textured mosaic sets, each composed of 100 grey or color mosaics made up of Brodatz or VisTex textures. Lastly, when applied to remote sensing images, our method yields good region segmentation compared to the ENVI commercial software, which demonstrates that the method could be used to generate land cover maps and is suitable for various purposes in image segmentation.     

迁移学习在医学图像分类中的研究进展

医学影像作为医疗数据的主要载体,在疾病预防、诊断和治疗中发挥着重要作用。医学图像分类是医学影像分析的重要组成部分。如何提高医学图像分类效率是一个持续的研究问题。随着计算机技术进步,医学图像分类方法已经从传统方法转到深度学习,再到目前热门的迁移学习。虽然迁移学习在医学图像分类中得到较广泛应用,但存在不少问题,本文对该领域的迁移学习应用情况进行综述,从中总结经验和发现问题,为未来研究提供线索。1)对基于迁移学习的医学图像分类研究的重要文献进行梳理、分析和总结,概括出3种迁移学习策略,即迁移模型的结构调整策略、参数调整策略和从迁移模型中提取特征的策略;2)从各文献研究设计的迁移学习过程中提炼共性,总结为5种迁移学习模式,即深度卷积神经网络(deep convolution neural network, DCNN)模式、混合模式、特征组合分类模式、多分类器融合模式和二次迁移模式。阐述了迁移学习策略和迁移学习模式之间的关系。这些迁移学习策略和模式有助于从更高的抽象层次展现迁移学习应用于医学图像分类领域的情况;3)阐述这些迁移学习策略和模式在医学图像分类中的具体应用,分析这些策略及模式的优点、局...     

面向元余弦损失的少样本图像分类

目的 度量学习是少样本学习中一种简单且有效的方法,学习一个丰富、具有判别性和泛化性强的嵌入空间是度量学习方法实现优秀分类效果的关键。本文从样本自身的特征以及特征在嵌入空间中的分布出发,结合全局与局部数据增强实现了一种元余弦损失的少样本图像分类方法(a meta-cosine loss for few-shot image classification,AMCL-FSIC)。方法 首先,从数据自身特征出发,将全局与局部的数据增广方法结合起来,利于局部信息提供更具区别性和迁移性的信息,使训练模型更多关注图像的前景信息。同时,利用注意力机制结合全局与局部特征,以得到更丰富更具判别性的特征。其次,从样本特征在嵌入空间中的分布出发,提出一种元余弦损失(meta-cosine loss,MCL)函数,优化少样本图像分类模型。使用样本与类原型间相似性的差调整不同类的原型,扩大类间距,使模型测试新任务时类间距更加明显,提升模型的泛化能力。结果 分别在5个少样本经典数据集上进行了实验对比,在FC100(Few-shot Cifar100)和CUB(Caltech-UCSD Birds-200-2011)数据集上,本文方法均达到了目前最优分类效果;在MiniImageNet、TieredImageNet和Cifar100数据集上与对比模型的结果相当。同时,在MiniImageNet,CUB和Cifar100数据集上进行对比实验以验证MCL的有效性,结果证明提出的MCL提升了余弦分类器的分类效果。结论 本文方法能充分提取少样本图像分类任务中的图像特征,有效提升度量学习在少样本图像分类中的准确率。     

Improvements in image categorization using codebook ensembles

The problem of object category classification by committees or ensembles of classifiers, each of which is based on one diverse codebook, is addressed in this paper. Two methods of constructing visual codebook ensembles are proposed in this study. The first technique introduces diverse individual visual codebooks using different clustering algorithms. The second uses various visual codebooks of different sizes for constructing an ensemble with high diversity. Codebook ensembles are trained to capture and convey image properties from different aspects. Based on these codebook ensembles, different types of image representations can be acquired. A classifier ensemble can be trained based on different expression datasets from the same training image set. The use of a classifier ensemble to categorize new images can lead to improved performance. Detailed experimental analysis on a Pascal VOC challenge dataset reveals that the present ensemble approach performs well, consistently improves the performance of visual object classifiers, and results in state-of-the-art performance in categorization.     

The search for optimal feature set in power quality event classification

The significance of detection and classification of power quality (PQ) events that disturbs the voltage and/or current waveforms in the electrical power distribution networks is well known. Consequently, in spite of a large number of research reports in this area, the problem of PQ event classification remains to be an important engineering problem. Several feature construction, pattern recognition, analysis, and classification methods were proposed for this purpose. In spite of the extensive number of such alternatives, a research on the comparison of “how useful these features with respect to each other using specific classifiers” was omitted. In this work, a thorough analysis is carried out regarding the classification strengths of an ensemble of celebrated features. The feature items were selected from well-known tools such as spectral information, wavelet extrema across several decomposition levels, and local statistical variations of the waveform. The tests are repeated for classification of several types of real-life data acquired during line-to-ground arcing faults and voltage sags due to the induction motor starting under different load conditions. In order to avoid specificity in classifier strength determination, eight different approaches are applied, including the computationally costly “exhaustive search” together with the leave-one-out technique. To further avoid specificity of the feature for a given classifier, two classifiers (Bayes and SVM) are tested. As a result of these analyses, the more useful set among a wider set of features for each classifier is obtained. It is observed that classification accuracy improves by eliminating relatively useless feature items for both classifiers. Furthermore, the feature selection results somewhat change according to the classifier used. This observation shows that when a new analysis tool or a feature is developed and claimed to perform “better” than another, one should always indicate the matching classifier for the feature because that feature may prove comparably inefficient with other classifiers.     

On the classification of remote sensing high spatial resolution image data

In this paper is investigated a methodology implementing an object-based approach to digital image classification using spectral and spatial attributes in a multiple-stage classifier structured as a binary tree. It is a well-established fact that object-based image classification is particularly appropriate when dealing with high spatial resolution image data. Following this approach, the image is initially segmented into objects that carry informational value. Next, spectral and spatial attributes are extracted from every object in the scene, and implemented into a classifier to produce a thematic map. As the combined number of spectral and spatial variables may become large compared to the number of available training samples, a reduction in the data dimensionality may be required whenever parametric classifiers are used, in order to mitigate the effects of the Hughes phenomenon. To this end the sequential feature selection (SFS) procedure is applied in a multiple-stage classifier structured as a binary tree. The advantage of a binary tree classifier lies in the fact that only one pair of classes is considered at each stage (node), allowing for an optimal selection of features. This proposed approach was tested using Quickbird image data covering an urban scene. The results are compared against results yielded by the traditional single-stage Gaussian maximum likelihood classifier. The results suggest the proposed methodology is adequate in the classification of high spatial resolution image data.     

A comparison of multiple classifier combinations using different voting-weights for remote sensing image classification

Remote sensing image classification is a common application of remote sensing images. In order to improve the performance of Remote sensing image classification, multiple classifier combinations are used to classify the Landsat-8 Operational Land Imager (Landsat-8 OLI) images. Some techniques and classifier combination algorithms are investigated. The classifier ensemble consisting of five member classifiers is constructed. The results of every member classifier are evaluated. The voting strategy is experimented to combine the classification results of the member classifier. The results show that all the classifiers have different performances and the multiple classifier combination provides better performance than a single classifier, and achieves higher overall accuracy of classification. The experiment shows that the multiple classifier combination using producer’s accuracy as voting-weight (MCC_mod2 and MCC_mod3) present higher classification accuracy than the algorithm using overall accuracy as voting-weight (MCC_mod1).And the multiple classifier combinations using different voting-weights affected the classification result in different land-cover types. The multiple classifier combination algorithm presented in this article using voting-weight based on the accuracy of multiple classifier may have stability problems, which need to be addressed in future studies.