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.     

Hybrid active learning for reducing the annotation effort of operators in classification systems

Active learning is understood as any form of learning in which the learning algorithm has some control over the input samples due to a specific sample selection process based on which it builds up the model. In this paper, we propose a novel active learning strategy for data-driven classifiers, which is based on unsupervised criterion during off-line training phase, followed by a supervised certainty-based criterion during incremental on-line training. In this sense, we call the new strategy hybrid active learning. Sample selection in the first phase is conducted from scratch (i.e. no initial labels/learners are needed) based on purely unsupervised criteria obtained from clusters: samples lying near cluster centers and near the borders of clusters are expected to represent the most informative ones regarding the distribution characteristics of the classes. In the second phase, the task is to update already trained classifiers during on-line mode with the most important samples in order to dynamically guide the classifier to more predictive power. Both strategies are essential for reducing the annotation and supervision effort of operators in off-line and on-line classification systems, as operators only have to label an exquisite subset of the off-line training data resp. give feedback only on specific occasions during on-line phase. The new active learning strategy is evaluated based on real-world data sets from UCI repository and collected at on-line quality control systems. The results show that an active learning based selection of training samples (1) does not weaken the classification accuracies compared to when using all samples in the training process and (2) can out-perform classifiers which are built on randomly selected data samples.     

Word segmentation of handwritten text using supervised classification techniques

Recent work on extracting features of gaps in handwritten text allows a classification of these gaps into inter-word and intra-word classes using suitable classification techniques. In this paper, we first analyse the features of the gaps using mutual information. We then investigate the underlying data distribution by using visualisation methods. These suggest that a complicated structure exists, which makes them difficult to be separated into two distinct classes. We apply five different supervised classification algorithms from the machine learning field on both the original dataset and a dataset with the best features selected using mutual information. Moreover, we improve the classification result with the aid of a set of feature variables of strokes preceding and following each gap. The classifiers are compared by employing McNemar's test. We find that SVMs and MLPs outperform the other classifiers and that preprocessing to select features works well. The best classification result attained suggests that the technique we employ is particularly suitable for digital ink manipulation at the level of words.     

Machine learning approach for automated visual inspection of machine components

Visual inspection on the surface of components is a main application of machine vision. Visual inspection finds its application in identifying defects such as scratches, cracks bubbles and measurement of cutting tool wear and welding quality. Machine learning approach to machine vision helps in automating the design process of machine vision systems. This approach involves image acquisition, preprocessing, feature extraction and classification. Study shows a library of features, and classifiers are available to classify the data. However, only the best combination of them can yield the highest classification accuracy. In this study, images with different known conditions were acquired, preprocessed, and histogram features were extracted. The classification accuracies of C4.5 classifier algorithm and Naïve Bayes algorithm were compared, and results are reported. The study shows that C4.5 algorithm performs better.     

Multitemporal ensemble learning for snow cover extraction from high-spatial-resolution images in mountain areas

ABSTRACT

High-spatial and -temporal resolution snow cover products in mountain areas are important to hydrological applications. The GF-1 satellite provides multispectral images with 8-m resolution and a revisit up to 2 days, which makes it possible to produce snow cover products. However, it is challenging to extract snow cover from these images because of limited spectral bands, severe mountain shadows, and dataset-shift problem in multitemporal classification. To overcome the limitations above, this study proposes a multitemporal ensemble learning framework to extract snow cover from high-spatial-resolution images in mountain areas. The principle behind ensemble learning, i.e. learning from disagreement, is extended from single image classification to multitemporal ones. We assume that multitemporal training samples selected within time-invariant classes at the same locations can be different in feature space. Such disagreements are used in multitemporal ensemble learning to improve classification accuracy. To enhance both accuracy and diversity of the multiple classifiers trained on these samples, a joint feature selection method is suggested to select the optimal multitemporal feature space and a joint parameter optimization method is designed to ensemble classifiers trained for multitemporal images. The experiments show that the performances of multitemporal ensemble classifiers are superior to that of single classifiers, confirming the effectiveness of the proposed framework.     

Diversity-aware classifier ensemble selection via f-score

The primary effect of using a reduced number of classifiers is a reduction in the computational requirements during learning and classification time. In addition to this obvious result, research shows that the fusion of all available classifiers is not a guarantee of best performance but good results on the average. The much researched issue of whether it is more convenient to fuse or to select has become even more of interest in recent years with the development of the Online Boosting theory, where a limited set of classifiers is continuously updated as new inputs are observed and classifications performed. The concept of online classification has recently received significant interest in the computer vision community. Classifiers can be trained on the visual features of a target, casting the tracking problem into a binary classification one: distinguishing the target from the background.Here we discuss how to optimize the performance of a classifier ensemble employed for target tracking in video sequences. In particular, we propose the F-score measure as a novel means to select the members of the ensemble in a dynamic fashion. For each frame, the ensemble is built as a subset of a larger pool of classifiers selecting its members according to their F-score. We observed an overall increase in classification accuracy and a general tendency in redundancy reduction among the members of an f-score optimized ensemble. We carried out our experiments both on benchmark binary datasets and standard video sequences.     

基于非线性降维多项式逻辑斯蒂回归的图像/非图像数据的分类与识别

在面向大规模复杂数据的模式分类和识别问题中,绝大多数的分类器都遇到了维数灾难这一棘手的问题.在进行高维数据分类之前,基于监督流形学习的非线性降维方法可提供一种有效的解决方法.利用多项式逻辑斯蒂回归方法进行分类预测,并结合基于非线性降维的非监督流形学习方法解决图像以及非图像数据的分类问题,因而形成了一种新的分类识别方法.大量的实验测试和比较分析验证了本文所提方法的优越性.     

基于非线性降维多项式逻辑斯蒂回归的图像/非图像数据的分类与识别(英文)

在面向大规模复杂数据的模式分类和识别问题中,绝大多数的分类器都遇到了维数灾难这一棘手的问题.在进行高维数据分类之前,基于监督流形学习的非线性降维方法可提供一种有效的解决方法.利用多项式逻辑斯蒂回归方法进行分类预测,并结合基于非线性降维的非监督流形学习方法解决图像以及非图像数据的分类问题,因而形成了一种新的分类识别方法.大量的实验测试和比较分析验证了本文所提方法的优越性.     

Identifying High Level Features of Texture Perception

A fundamental issue in texture analysis is that of deciding what textural features are important in texture perception, and how they are used. Experiments on human preattentive vision have identified several low-level features (such as orientation of blobs and size of line segments), which are used in texture perception. However, the question of what higher level features of texture are used has not been adequately addressed. We designed an experiment to help identify the relevant higher order features of texture perceived by humans. We used 20 subjects, who were asked to perform an unsupervised classification of 30 pictures from Brodatz′s album on texture. Each subject was asked to group these pictures into as many classes as desired. Both hierarchical cluster analysis and nonparametric multidimensional scaling (MDS) were applied to the pooled similarity matrix generated from the subjects′ groupings. A surprising outcome is that the MDS solutions fit the data very well. The stress in the two-dimensional case is 0.10, and the stress in the three-dimensional case is 0.045. We rendered the original textures in these coordinate systems, and interpreted the (rotated) axes. It appears that the axes in the 2D case correspond to periodicity versus irregularity, and directionality versus nondirectionality. In the 3D case, the third dimension represents the structural complexity of the texture. Furthermore, the clusters identified by the hierarchical cluster analysis remain virtually intact in the MDS solution. The results of our experiment indicate that people use three high-level features for texture perception. Future studies are needed to determine the appropriateness of these high-level features for computational texture analysis and classification.     

多类不平衡数据分类方法综述

现实中许多领域产生的数据通常具有多个类别并且是不平衡的。在多类不平衡分类中,类重叠、噪声和多个少数类等问题降低了分类器的能力,而有效解决多类不平衡问题已经成为机器学习与数据挖掘领域中重要的研究课题。根据近年来的多类不平衡分类方法的文献,从数据预处理和算法级分类方法两方面进行了分析与总结,并从优缺点和数据集等方面对所有算法进行了详细的分析。在数据预处理方法中,介绍了过采样、欠采样、混合采样和特征选择方法,对使用相同数据集算法的性能进行了比较。从基分类器优化、集成学习和多类分解技术三个方面对算法级分类方法展开介绍和分析。最后对多类不平衡数据分类研究领域的未来发展方向进行总结归纳。     

Discretization for naive-Bayes learning: managing discretization bias and variance

Quantitative attributes are usually discretized in Naive-Bayes learning. We establish simple conditions under which discretization is equivalent to use of the true probability density function during naive-Bayes learning. The use of different discretization techniques can be expected to affect the classification bias and variance of generated naive-Bayes classifiers, effects we name discretization bias and variance. We argue that by properly managing discretization bias and variance, we can effectively reduce naive-Bayes classification error. In particular, we supply insights into managing discretization bias and variance by adjusting the number of intervals and the number of training instances contained in each interval. We accordingly propose proportional discretization and fixed frequency discretization, two efficient unsupervised discretization methods that are able to effectively manage discretization bias and variance. We evaluate our new techniques against four key discretization methods for naive-Bayes classifiers. The experimental results support our theoretical analyses by showing that with statistically significant frequency, naive-Bayes classifiers trained on data discretized by our new methods are able to achieve lower classification error than those trained on data discretized by current established discretization methods.     

Texture classification using partial differential equation approach and wavelet transform

Textures and patterns are the distinguishing characteristics of objects. Texture classification plays fundamental role in computer vision and image processing applications. In this paper, texture classification using PDE (partial differential equation) approach and wavelet transform is presented. The proposed method uses wavelet transform to obtain the directional information of the image. A PDE for anisotropic diffusion is employed to obtain texture component of the image. The feature set is obtained by computing different statistical features from the texture component. The linear discriminant analysis (LDA) enhances separability of texture feature classes. The features obtained from LDA are class representatives. The proposed approach is experimented on three gray scale texture datasets: VisTex, Kylberg, and Oulu. The classification accuracy of the proposed method is evaluated using k-NN classifier. The experimental results show the effectiveness of the proposed method as compared to the other methods in the literature.     

Multi-class supervised classification of electrical borehole wall images using texture features

Electrical borehole wall images represent micro-resistivity measurements at the borehole wall. The lithology reconstruction is often based on visual interpretation done by geologists. This analysis is very time-consuming and subjective. Different geologists may interpret the data differently. In this work, linear discriminant analysis (LDA) in combination with texture features is used for an automated lithology reconstruction of ODP (Ocean Drilling Program) borehole 1203A drilled during Leg 197. Six rock groups are identified by their textural properties in resistivity data obtained by a Formation MircoScanner (FMS). Although discriminant analysis can be used for multi-class classification, non-optimal decision criteria for certain groups could emerge. For this reason, we use a combination of 2-class (binary) classifiers to increase the overall classification accuracy. The generalization ability of the combined classifiers is evaluated and optimized on a testing dataset where a classification rate of more than 80% for each of the six rock groups is achieved. The combined, trained classifiers are then applied on the whole dataset obtaining a statistical reconstruction of the logged formation. Compared to a single multi-class classifier the combined binary classifiers show better classification results for certain rock groups and more stable results in larger intervals of equal rock type.     

Optimised deep learning features for improved melanoma detection

In this article, we are addressing the question of effective usage of the feature set extracted from deep learning models pre-trained on ImageNet. Exploring this option will offer very fast and attractive alternative to transfer learning strategies. The traditional task of skin lesion recognition consists of several stages, where the automated system is typically trained on preprocessed images with known diagnosis, which allows classification of new samples to predefined categories. For this task, we are proposing here an improved melanoma detection method based on the combination of linear discriminant analysis (LDA) and the features extracted from the deep learning approach. We are examining the usage of the LDA approach on activation of the fully-connected layer of deep learning in order to increase the classification accuracy and at the same time to reduce the feature space dimensionality. We tested our method on five different classifiers and evaluated results using various metrics. The presented comparison demonstrates the very high effectiveness of the suggested feature reduction, which leads not only to the significant lowering of employed features but also to the increasing performance of all tested classifiers in almost all measured characteristics.

     

Visual domain adaptation via transfer feature learning

One of the serious challenges in computer vision and image classification is learning an accurate classifier for a new unlabeled image dataset, considering that there is no available labeled training data. Transfer learning and domain adaptation are two outstanding solutions that tackle this challenge by employing available datasets, even with significant difference in distribution and properties, and transfer the knowledge from a related domain to the target domain. The main difference between these two solutions is their primary assumption about change in marginal and conditional distributions where transfer learning emphasizes on problems with same marginal distribution and different conditional distribution, and domain adaptation deals with opposite conditions. Most prior works have exploited these two learning strategies separately for domain shift problem where training and test sets are drawn from different distributions. In this paper, we exploit joint transfer learning and domain adaptation to cope with domain shift problem in which the distribution difference is significantly large, particularly vision datasets. We therefore put forward a novel transfer learning and domain adaptation approach, referred to as visual domain adaptation (VDA). Specifically, VDA reduces the joint marginal and conditional distributions across domains in an unsupervised manner where no label is available in test set. Moreover, VDA constructs condensed domain invariant clusters in the embedding representation to separate various classes alongside the domain transfer. In this work, we employ pseudo target labels refinement to iteratively converge to final solution. Employing an iterative procedure along with a novel optimization problem creates a robust and effective representation for adaptation across domains. Extensive experiments on 16 real vision datasets with different difficulties verify that VDA can significantly outperform state-of-the-art methods in image classification problem.     

Multi-class support vector machine for classification of the ultrasonic images of supraspinatus

This article proposes an effort to apply the multi-class support vector machine classifiers to classify the supraspinatus image into different disease groups that are normal, tendon inflammation, calcific tendonitis and supraspinatus tear. The supraspinatus tendon is often involved in the above-mentioned disease groups. Four different texture analysis methods – texture feature coding method, gray-level co-occurrence matrix, fractal dimension evaluation and texture spectrum – are used to extract features of tissue characteristic in the ultrasonic supraspinatus images. The mutual information criterion is adopted to select the powerful features from ones generated from the above-mentioned four texture analysis methods in the training stage, meanwhile, the five implementations of multi-class support vector machine classifiers are also designed to discriminate each image into one of the four disease groups in the classification stage. In experiments, the most commonly used performance measures including sensitivity, specificity, classification accuracy and false-negative rate are applied to evaluate the classification of the five implantations of multi-class support vector machines. In addition, the receiver operating characteristics analysis is also used to analyze the classification capability. The present results demonstrate that the implementation of multi-class fuzzy support vector machine can achieve 90% classification accuracy, and performance measures of this implementation are significantly superior to the others.     

稀疏降噪自编码器在IR-BCI的应用研究

针对脑-机接口的特征提取问题,提出了一种基于非监督学习的稀疏降噪自编码器,对刺激诱发的脑电信号进行自主学习,构建原始数据的深层特征表达。该编码器引用稀疏自编码神经网络,通过加入噪声,增强其学习的泛化能力,增加了神经网络的鲁棒性。首先对多导联信号进行重新拼接,输入稀疏降噪自编码器,得到原始数据的稀疏特征表达;然后,采用支持向量机将学习到的特征进行分类;最后,同直接使用最优单通道相对比。实验结果为：稀疏降噪自编码器的分类准确率要优于单通道,表明该方法能够更好地学习到特征,并提高了“模拟阅读”脑-机接口的识别正确率,为脑-机接口系统的特征提取和分类提供了新思路。     

Unsupervised spectral sub-feature learning for hyperspectral image classification

Spectral pixel classification is one of the principal techniques used in hyperspectral image (HSI) analysis. In this article, we propose an unsupervised feature learning method for classification of hyperspectral images. The proposed method learns a dictionary of sub-feature basis representations from the spectral domain, which allows effective use of the correlated spectral data. The learned dictionary is then used in encoding convolutional samples from the hyperspectral input pixels to an expanded but sparse feature space. Expanded hyperspectral feature representations enable linear separation between object classes present in an image. To evaluate the proposed method, we performed experiments on several commonly used HSI data sets acquired at different locations and by different sensors. Our experimental results show that the proposed method outperforms other pixel-wise classification methods that make use of unsupervised feature extraction approaches. Additionally, even though our approach does not use any prior knowledge, or labelled training data to learn features, it yields either advantageous, or comparable, results in terms of classification accuracy with respect to recent semi-supervised methods.