Cloud classification from visible and infrared SMS-1 data

The success of a statistical classification-design sample model in discriminating cloud-type samples of visible and infrared meteorological satellite data depends on the selection of the design parameters for the system and the ability of the labelled design samples to characterize and discriminate class patterns within the given geographical region. In a companion study by Parikh (1977), pattern recognition design parameters were examined for a four-class problem and a three-class problem for NOAA-1 cloud data. The purpose of this study was to evaluate pattern recognition systems designed in the previous study on SMS-1 design and test sets. Experiments were conducted for both a four-class problem (separation of “low”, “mix”, “cirrus”, and “cumulonimbus” samples) and a three-class problem (separation of “low”, “cirrus”, and “cumulonimbus” samples). For the four-class problem, decreases in classification accuracy ranging from 4% to 11% occurred when the pattern recognition systems were designed and tested on two different data sets selected from the same satellite orbit. A similar decrease was not observed for the well-defined three-class problem.     

Textural and spectral features as an aid to cloud classification

Abstract

The problem of classifying clouds seen on meteorological satellite images into different types is one which requires the use of textural as well as spectral information. Since multi-spectral features are of prime importance, textural features must be considered as augmenting, rather than replacing, spectral measures. Several textural features are studied to determine their discriminating power across a number of cloud classes including those which have previously been found difficult to separate. Although several features in the frequency domain are tested they are found to be less useful than those in the spatial domain with only one exception. The specific features recommended for use in classification depend on the type of classification to be undertaken. Specifically, different features should be used for a multi-dimensional feature space analysis than for a binary-tree rule-based classification.     

云的光谱和纹理特征统计分析

利用静止卫星图像资料建立了夏季白天中低纬地区的11 种云/ 表面类型的样本集, 从中随机 挑选656 个样本, 提取116 个光谱和纹理特征参数并进行统计分析, 通过特征选择组成特征向量, 带入逐个修改聚类和模糊聚类的分类器进行敏感性试验。结果发现, 在反映云特征方面, 光谱特征 是云分类最基本的特征, 比纹理特征明显, 是云分类识别的主要依据; 除去水汽通道的标准差以外 其它光谱特征都比较明显, 红外和水汽通道的特征明显好于可见光通道, 尤其是对中低云和卷云的 描述。纹理特征在反映云特征方面也有一定的代表性, 特别是一阶概率特征中四通道的惯量及水汽 通道的逆差距; 纹理特征引入后分类准确率显著提高, 但在引入一阶概率特征基础上引入灰度级差 矢量特征效果改善并不明显。     

ON IMAGE CLASSIFICATION: CITY IMAGES VS. LANDSCAPES

Grouping images into semantically meaningful categories using low-level visual features is a challenging and important problem in content-based image retrieval. Based on these groupings, effective indices can be built for an image database. In this paper, we show how a specific high-level classification problem (city images vs landscapes) can be solved from relatively simple low-level features geared for the particular classes. We have developed a procedure to qualitatively measure the saliency of a feature towards a classification problem based on the plot of the intra-class and inter-class distance distributions. We use this approach to determine the discriminative power of the following features: color histogram, color coherence vector, DCT coefficient, edge direction histogram, and edge direction coherence vector. We determine that the edge direction-based features have the most discriminative power for the classification problem of interest here. A weighted k-NN classifier is used for the classification which results in an accuracy of 93.9% when evaluated on an image database of 2716 images using the leave-one-out method. This approach has been extended to further classify 528 landscape images into forests, mountains, and sunset/sunrise classes. First, the input images are classified as sunset/sunrise images vs forest & mountain images (94.5% accuracy) and then the forest & mountain images are classified as forest images or mountain images (91.7% accuracy). We are currently identifying further semantic classes to assign to images as well as extracting low level features which are salient for these classes. Our final goal is to combine multiple 2-class classifiers into a single hierarchical classifier.     

Handwriting style classification

This paper describes an independent handwriting style classifier that has been designed to select the best recognizer for a given style of writing. For this purpose a definition of handwriting legibility has been defined and a method implemented that can predict this legibility. The technique consists of two phases. In the feature-extraction phase, a set of 36 features is extracted from the image contour. In the classification phase, two nonparametric classification techniques are applied to the extracted features in order to compare their effectiveness in classifying words into legible, illegible, and middle classes. In the first method, a multiple discriminant analysis (MDA) is used to transform the space of extracted features (36 dimensions) into an optimal discriminant space for a nearest mean based classifier. In the second method, a probabilistic neural network (PNN) based on the Bayes strategy and nonparametric estimation of probability density function is used. The experimental results show that the PNN method gives superior classification results when compared with the MDA method. For the legible, illegible, and middle handwriting the method provides 86.5% (legible/illegible), 65.5% (legible/middle), and 90.5% (middle/illegible) correct classification for two classes. For the three-class legibility classification the rate of correct classification is 67.33% using a PNN classifier.Received: 6 September 2002, Accepted: 19 September 2002, Published online: 6 June 2003     

A comparison of hierarchical multi-output recognition approaches for anuran classification

In bioacoustic recognition approaches, a “flat” classifier is usually trained to recognize several species of anurans, where the number of classes is equal to the number of species. Consequently, the complexity of the classification function increases proportionally with the number of species. To avoid this issue, we propose a “hierarchical” approach that decomposes the problem into three taxonomic levels: the family, the genus, and the species. To accomplish this, we transform the original single-labelled problem into a multi-output problem (multi-label and multi-class) considering the biological taxonomy of the species. We then develop a top-down method using a set of classifiers organized as a hierarchical tree. We test and compare two hierarchical methods, using (1) one classifier per parent node and (2) one classifier per level, against a flat approach. Thus, we conclude that it is possible to predict the same set of species as a flat classifier, and additionally obtain new information about the samples and their taxonomic relationship. This helps us to better understand the problem and achieve additional conclusions by the inspection of the confusion matrices at the three classification levels. In addition, we propose a soft decision rule based on the joint probabilities of hierarchy pathways. With this we are able to identify and reject confusing cases. We carry out our experiments using cross-validation performed by individuals. This form of CV avoids mixing syllables that belong to the same specimens in the testing and training sets, preventing an overestimate of the accuracy and generalizing the predictive capabilities of the system. We tested our methods in a dataset with sixty individual frogs, from ten different species, eight genera, and four families, achieving a final Macro-Fscore of 80 and 70% with and without applying the rejection rule, respectively.     

A study of cloud classification with neural networks using spectraland textural features

The problem of cloud data classification from satellite imagery using neural networks is considered. Several image transformations such as singular value decomposition (SVD) and wavelet packet (WP) were used to extract the salient spectral and textural features attributed to satellite cloud data in both visible and infrared (IR) channels. In addition, the well-known gray-level cooccurrence matrix (GLCM) method and spectral features were examined for the sake of comparison. Two different neural-network paradigms namely probability neural network (PNN) and unsupervised Kohonen self-organized feature map (SOM) were examined and their performance were also benchmarked on the geostationary operational environmental satellite (GOES) 8 data. Additionally, a postprocessing scheme was developed which utilizes the contextual information in the satellite images to improve the final classification accuracy. Overall, the performance of the PNN when used in conjunction with these feature extraction and postprocessing schemes showed the potential of this neural-network-based cloud classification system.     

Fingerprint classification by a hierarchical classifier

Fingerprint classification is still a challenging problem due to large intra-class variability, small inter-class variability and the presence of noise. To deal with these difficulties, we propose a regularized orientation diffusion model for fingerprint orientation extraction and a hierarchical classifier for fingerprint classification in this paper. The proposed classification algorithm is composed of five cascading stages. The first stage rapidly distinguishes a majority of Arch by using complex filter responses. The second stage distinguishes a majority of Whorl by using core points and ridge line flow classifier. In the third stage, K-NN classifier finds the top two categories by using orientation field and complex filter responses. In the fourth stage, ridge line flow classifier is used to distinguish Loop from other classes except Whorl. SVM is adopted to make the final classification in the last stage. The regularized orientation diffusion model has been evaluated on a web-based automated evaluation system FVC-onGoing, and a promising result is obtained. The classification method has been evaluated on the NIST SD 4. It achieved a classification accuracy of 95.9% for five-class classification and 97.2% for four-class classification without rejection.     

A supervised learning algorithm for hierarchical classification of fuzzy patterns

The notion of “fuzzy separability” is introduced for fuzzy sets of patterns. A supervised learning algorithm is proposed for estimation of membership functions that yield hierarchical partitioning of the feature space for fuzzy separable pattern classes under confusion. Finally we present a methodology for the design of a classifier composed of hierarchical binary decision trees.     

Analysis of single- and dual-dictionary strategies in pedestrian classification

Sparse coding has recently been a hot topic in visual tasks in image processing and computer vision. It has applications and brings benefits in reconstruction-like tasks and in classification-like tasks as well. However, regarding binary classification problems, there are several choices to learn and use dictionaries that have not been studied. In particular, how single-dictionary and dual-dictionary approaches compare in terms of classification performance is largely unexplored. We compare three single-dictionary strategies and two dual-dictionary strategies for the problem of pedestrian classification (“pedestrian” vs “background” images). In each of these five cases, images are represented as the sparse coefficients induced from the respective dictionaries, and these coefficients are the input to a regular classifier both for training and subsequent classification of novel unseen instances. Experimental results with the INRIA pedestrian dataset suggest, on the one hand, that dictionaries learned from only one of the classes, even from the background class, are enough for obtaining competitive good classification performance. On the other hand, while better performance is generally obtained when instances of both classes are used for dictionary learning, the representation induced by a single dictionary learned from a set of instances from both classes provides comparable or even superior performance over the representations induced by two dictionaries learned separately from the pedestrian and background classes.     

Can computer vision problems benefit from structured hierarchical classification?

Research in the field of supervised classification has mostly focused on the standard, so-called “flat” classification approach, where the problem classes live in a trivial, one-level semantic space. There is however an increasing interest in the hierarchical classification approach, where a performance gain is expected by incorporating prior taxonomic knowledge about the classes into the learning process. Intuitively, the hierarchical approach should be beneficial in general for the classification of visual content, as suggested by the fact that humans seem to organize objects into hierarchies based on visually perceived similarities. In this paper, we provide an analysis that aims to determine the conditions under which the hierarchical approach can consistently give better performances than the flat approach for the classification of visual content. In particular, we (1) show how hierarchical methods can fail to outperform flat methods when applied to real vision-based classification problems, and (2) investigate the underlying reasons for the lack of improvement, by applying the same methods to synthetic datasets in a simulation. Our conclusion is that the use of high-level hierarchical feature representations is crucial for obtaining a performance gain with the hierarchical approach, and that poorly chosen prior taxonomies hinder this gain even though proper high-level features are used.     

Texture classification of logged forests in tropical Africa using machine-learning algorithms

This Letter describes a procedure that incorporates textural measures in the classification of logged forests from Landsat Thematic Mapper data. The objective was to increase classification accuracy by applying recently developed algorithms in machine learning that are fast in training. Three voting classification algorithms, Arc-4x, Adaboost and bagging were also tested. Initial results using a decision tree classifier showed that adding selected textural measures increased the accuracy of logged forest classification by almost 40%, although the class accuracy for logged forests was only approximately 50% when using spectral and textural features combined. No further significant increase in the classification of logged forests was obtained by voting classification.     

Object-based urban detailed land cover classification with high spatial resolution IKONOS imagery

Improvement in remote sensing techniques in spatial/spectral resolution strengthens their applicability for urban environmental study. Unfortunately, high spatial resolution imagery also increases internal variability in land cover units and can cause a ‘salt-and-pepper’ effect, resulting in decreased accuracy using pixel-based classification results. Region-based classification techniques, using an image object (IO) rather than a pixel as a classification unit, appear to hold promise as a method for overcoming this problem. Using IKONOS high spatial resolution imagery, we examined whether the IO technique could significantly improve classification accuracy compared to the pixel-based method when applied to urban land cover mapping in Tampa Bay, FL, USA. We further compared the performance of an artificial neural network (ANN) and a minimum distance classifier (MDC) in urban detailed land cover classification and evaluated whether the classification accuracy was affected by the number of extracted IO features. Our analysis methods included IKONOS image data calibration, data fusion with the pansharpening (PS) process, Hue–Intensity–Saturation (HIS) transferred indices and textural feature extraction, and feature selection using a stepwise discriminant analysis (SDA). The classification results were evaluated with visually interpreted data from high-resolution (0.3 m) digital aerial photographs. Our results indicate a statistically significant difference in classification accuracy between pixel- and object-based techniques; ANN outperforms MDC as an object-based classifier; and the use of more features (27 vs. 9 features) increases the IO classification accuracy, although the increase is statistically significant for the MDC but not for the ANN.     

基于ReliefF的层次分类在线流特征选择算法

在图像标注、疾病诊断等实际分类任务中,数据标记空间的类别通常存在着层次化结构关系,且伴随着特征的高维性.许多层次特征选择算法因不同的实际任务需求而提出,但这些已有的特征选择算法忽略了特征空间的未知性和不确定性.针对上述问题,提出一种基于ReliefF的面向层次分类学习的在线流特征选择算法OH_ReliefF.首先将类别...     

多级中文文本情感分类算法研究

针对文本情感分类准确率不高的问题,提出基于CCA-VSM分类器和KFD的多级文本情感分类方法。采用典型相关性分析对文档的权重特征向量和词性特征向量进行降维,在约简向量集上构建向量空间模型,根据模型之间的差异度设计VSM分类器,筛选出与测试文档差异度较小的R个模型作为核Fisher判别的输入,最终判别出文档的情感观点。实验结果表明:该方法比传统支持向量机有较高的分类准确率和较快的分类速度,权重特征和词性特征对分类准确率的影响较大。     

联机手写维吾尔文字母识别方法

通过分析维吾尔文字母自身的结构和书写特点,提出一种联机手写维吾尔文字母识别方案,并选择在手写汉字识别技术中所提出来的归一化、特征提取及常用的分类方法,从中找出最佳的技术选择。在实验对比中,采用8种不同的归一化预处理方法,基于坐标归一化的特征提取 (NCFE) 方法,以及改进的二次分类函数(MQDF)、判别学习型二次判别函数(DLQDF)、学习矢量量化(LVQ)、支持向量机(SVM)4种分类器。同时,再考虑字符在文档中的空间几何特征,进一步提高识别性能。在128个维吾尔文字母类别、38 400个测试样本的实验中,正确识别率最高达89。08%,为进一步研究面向维吾尔文字母特性的识别技术奠定重要基础。     

Improving handwriting based gender classification using ensemble classifiers

This paper presents a system to predict gender of individuals from offline handwriting samples. The technique relies on extracting a set of textural features from handwriting samples of male and female writers and training multiple classifiers to learn to discriminate between the two gender classes. The features include local binary patterns (LBP), histogram of oriented gradients (HOG), statistics computed from gray-level co-occurrence matrices (GLCM) and features extracted through segmentation-based fractal texture analysis (SFTA). For classification, we employ artificial neural networks (ANN), support vector machine (SVM), nearest neighbor classifier (NN), decision trees (DT) and random forests (RF). Classifiers are then combined using bagging, voting and stacking techniques to enhance the overall system performance. The realized classification rates are significantly better than those of the state-of-the-art systems on this problem validating the ideas put forward in this study.     

ClassView: hierarchical video shot classification, indexing, and accessing

Recent advances in digital video compression and networks have made video more accessible than ever. However, the existing content-based video retrieval systems still suffer from the following problems. 1) Semantics-sensitive video classification problem because of the semantic gap between low-level visual features and high-level semantic visual concepts; 2) Integrated video access problem because of the lack of efficient video database indexing, automatic video annotation, and concept-oriented summary organization techniques. In this paper, we have proposed a novel framework, called ClassView, to make some advances toward more efficient video database indexing and access. 1) A hierarchical semantics-sensitive video classifier is proposed to shorten the semantic gap. The hierarchical tree structure of the semantics-sensitive video classifier is derived from the domain-dependent concept hierarchy of video contents in a database. Relevance analysis is used for selecting the discriminating visual features with suitable importances. The Expectation-Maximization (EM) algorithm is also used to determine the classification rule for each visual concept node in the classifier. 2) A hierarchical video database indexing and summary presentation technique is proposed to support more effective video access over a large-scale database. The hierarchical tree structure of our video database indexing scheme is determined by the domain-dependent concept hierarchy which is also used for video classification. The presentation of visual summary is also integrated with the inherent hierarchical video database indexing tree structure. Integrating video access with efficient database indexing tree structure has provided great opportunity for supporting more powerful video search engines.     

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.     

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.