Showing papers on "Contextual image classification published in 2001"

Rapid object detection using a boosted cascade of simple features

Abstract: This paper describes a machine learning approach for visual object detection which is capable of processing images extremely rapidly and achieving high detection rates. This work is distinguished by three key contributions. The first is the introduction of a new image representation called the "integral image" which allows the features used by our detector to be computed very quickly. The second is a learning algorithm, based on AdaBoost, which selects a small number of critical visual features from a larger set and yields extremely efficient classifiers. The third contribution is a method for combining increasingly more complex classifiers in a "cascade" which allows background regions of the image to be quickly discarded while spending more computation on promising object-like regions. The cascade can be viewed as an object specific focus-of-attention mechanism which unlike previous approaches provides statistical guarantees that discarded regions are unlikely to contain the object of interest. In the domain of face detection the system yields detection rates comparable to the best previous systems. Used in real-time applications, the detector runs at 15 frames per second without resorting to image differencing or skin color detection.

SIMPLIcity: semantics-sensitive integrated matching for picture libraries

Abstract: We present here SIMPLIcity (semantics-sensitive integrated matching for picture libraries), an image retrieval system, which uses semantics classification methods, a wavelet-based approach for feature extraction, and integrated region matching based upon image segmentation. An image is represented by a set of regions, roughly corresponding to objects, which are characterized by color, texture, shape, and location. The system classifies images into semantic categories. Potentially, the categorization enhances retrieval by permitting semantically-adaptive searching methods and narrowing down the searching range in a database. A measure for the overall similarity between images is developed using a region-matching scheme that integrates properties of all the regions in the images. The application of SIMPLIcity to several databases has demonstrated that our system performs significantly better and faster than existing ones. The system is fairly robust to image alterations.

Image classification for content-based indexing

Abstract: Grouping images into (semantically) meaningful categories using low-level visual features is a challenging and important problem in content-based image retrieval. Using binary Bayesian classifiers, we attempt to capture high-level concepts from low-level image features under the constraint that the test image does belong to one of the classes. Specifically, we consider the hierarchical classification of vacation images; at the highest level, images are classified as indoor or outdoor; outdoor images are further classified as city or landscape; finally, a subset of landscape images is classified into sunset, forest, and mountain classes. We demonstrate that a small vector quantizer (whose optimal size is selected using a modified MDL criterion) can be used to model the class-conditional densities of the features, required by the Bayesian methodology. The classifiers have been designed and evaluated on a database of 6931 vacation photographs. Our system achieved a classification accuracy of 90.5% for indoor/outdoor, 95.3% for city/landscape, 96.6% for sunset/forest and mountain, and 96% for forest/mountain classification problems. We further develop a learning method to incrementally train the classifiers as additional data become available. We also show preliminary results for feature reduction using clustering techniques. Our goal is to combine multiple two-class classifiers into a single hierarchical classifier.

Classification Methods for Remotely Sensed Data

Abstract: Preface to the Second Edition Preface to the First Edition Author Biographies Chapter 1: Remote Sensing in the Optical and Microwave Regions 1.1 Introduction to Remote Sensing 1.1.1 Atmospheric Interactions 1.1.2 Surface Material Reflectance 1.1.3 Spatial and Radiometric Resolution 1.2 Optical Remote Sensing Systems 1.3 Atmospheric Correction 1.3.1 Dark Object Subtraction 1.3.2 Modeling Techniques 1.3.2.1 Modeling the Atmospheric Effect 1.3.2.2 Steps in Atmospheric Correction 1.4 Correction for Topographic Effects 1.5 Remote Sensing in the Microwave Region 1.6 Radar Fundamentals 1.6.1 SLAR Image Resolution 1.6.2 Geometric Effects on Radar Images 1.6.3 Factors Affecting Radar Backscatter 1.6.3.1 Surface Roughness 1.6.3.2 Surface Conductivity 1.6.3.3 Parameters of the Radar Equation 1.7 Imaging Radar Polarimetry 1.7.1 Radar Polarization State 1.7.2 Polarization Synthesis 1.7.3 Polarization Signatures 1.8 Radar Speckle Suppression 1.8.1 Multilook Processing 1.8.2 Filters for Speckle Suppression Chapter 2: Pattern Recognition Principles 2.1 Feature Space Manipulation 2.1.1 Tasseled Cap Transform 2.1.2 Principal Components Analysis 2.1.3 Minimum/Maximum Autocorrelation Factors (MAF) 2.1.4 Maximum Noise Fraction Transformation 2.2 Feature Selection 2.3 Fundamental Pattern Recognition Techniques 2.3.1 Unsupervised Methods 2.3.1.1 The k-means Algorithm 2.3.1.2 Fuzzy Clustering 2.3.2 Supervised Methods 2.3.2.1 Parallelepiped Method 2.3.2.2 Minimum Distance Classifier 2.3.2.3 Maximum Likelihood Classifier 2.4 Combining Classifiers 2.5 Incorporation of Ancillary Information 2.5.1 Use of Texture and Context 2.5.2 Using Ancillary Multisource Data 2.6 Sampling Scheme and Sample Size 2.6.1 Sampling Scheme 2.6.2 Sample Size, Scale, and Spatial Variability 2.6.3 Adequacy of Training Data 2.7 Estimation of Classification Accuracy Epilogue Chapter 3: Artificial Neural Networks 3.1 Multilayer Perceptron 3.1.1 Back-Propagation 3.1.2 Parameter Choice, Network Architecture, and Input/Output Coding 3.1.3 Decision Boundaries in Feature Space 3.1.4 Overtraining and Network Pruning 3.2 Kohonen's Self-Organizing Feature Map 3.2.1 SOM Network Construction and Training 3.2.1.1 Unsupervised Training 3.2.1.2 Supervised Training 3.2.2 Examples of Self-Organization 3.3 Counter-Propagation Networks 3.3.1 Counter-Propagation Network Training 3.3.2 Training Issues 3.4 Hopfield Networks 3.4.1 Hopfield Network Structure 3.4.2 Hopfield Network Dynamics 3.4.3 Network Convergence 3.4.4 Issues Relating to Hopfield Networks 3.4.5 Energy and Weight Coding: An Example 3.5 Adaptive Resonance Theory (ART) 3.5.1 Fundamentals of the ART Model 3.5.2 Choice of Parameters 3.5.3 Fuzzy ARTMAP 3.6 Neural Networks in Remote Sensing Image Classification 3.6.1 An Overview 3.6.2 A Comparative Study Chapter 4: Support Vector Machines 4.1 Linear Classification 4.1.1 The Separable Case4.1.2 The Nonseparable Case 4.2 Nonlinear Classification and Kernel Functions 4.2.1 Nonlinear SVMs 4.2.2 Kernel Functions 4.3 Parameter Determination 4.3.1 t-fold Cross-Validations 4.3.2 Bound on Leave-One-Out Error 4.3.3 Grid Search 4.3.4 Gradient Descent Method 4.4 Multiclass Classification 4.4.1 One-against-One, One-against-Others, and DAG 4.4.2 Multiclass SVMs 4.4.2.1 Vapnik's Approach 4.4.2.2 Methodology of Crammer and Singer 4.5 Feature Selection 4.6 SVM Classification of Remotely Sensed Data 4.7 Concluding Remarks Chapter 5: Methods Based on Fuzzy Set Theory 5.1 Introduction to Fuzzy Set Theory 5.1.1 Fuzzy Sets: Definition 5.1.2 Fuzzy Set Operations 5.2 Fuzzy C-Means Clustering Algorithm 5.3 Fuzzy Maximum Likelihood Classification 5.4 Fuzzy Rule Base 5.4.1 Fuzzification 5.4.2 Inference 5.4.3 Defuzzification 5.5 Image Classification Using Fuzzy Rules 5.5.1 Introductory Methodology 5.5.2 Experimental Results Chapter 6: Decision Trees 6.1 Feature Selection Measures for Tree Induction 6.1.1 Information Gain 6.1.2 Gini Impurity Index 6.2 ID3, C4.5, and SEE5.0 Decision Trees 6.2.1 ID3 6.2.2 C4.5 6.2.3 SEE5.0 6.3 CHAID 6.4 CART 6.5 QUEST 6.5.1 Split Point Selection 6.5.2 Attribute Selection 6.6 Tree Induction from Artificial Neural Networks 6.7 Pruning Decision Trees 6.7.1 Reduced Error Pruning (REP) 6.7.2 Pessimistic Error Pruning (PEP) 6.7.3 Error-Based Pruning (EBP) 6.7.4 Cost Complexity Pruning (CCP) 6.7.5 Minimal Error Pruning (MEP) 6.8 Boosting and Random Forest 6.8.1 Boosting 6.8.2 Random Forest 6.9 Decision Trees in Remotely Sensed Data Classification 6.10 Concluding Remarks Chapter 7: Texture Quantization 7.1 Fractal Dimensions 7.1.1 Introduction to Fractals 7.1.2 Estimation of the Fractal Dimension 7.1.2.1 Fractal Brownian Motion (FBM) 7.1.2.2 Box-Counting Methods and Multifractal Dimension 7.2 Frequency Domain Filtering 7.2.1 Fourier Power Spectrum 7.2.2 Wavelet Transform 7.3 Gray-Level Co-occurrence Matrix (GLCM) 7.3.1 Introduction to the GLCM 7.3.2 Texture Features Derived from the GLCM 7.4 Multiplicative Autoregressive Random Fields 7.4.1 MAR Model: Definition 7.4.2 Estimation of the Parameters of the MAR Model 7.5 The Semivariogram and Window Size Determination 7.6 Experimental Analysis 7.6.1 Test Image Generation 7.6.2 Choice of Texture Features 7.6.2.1 Multifractal Dimension 7.6.2.2 Fourier Power Spectrum 7.6.2.3 Wavelet Transform 7.6.2.4 Gray-Level Co-occurrence Matrix 7.6.2.5 Multiplicative Autoregressive Random Field 7.6.3 Segmentation Results 7.6.4 Texture Measure of Remote Sensing Patterns Chapter 8: Modeling Context Using Markov Random Fields 8.1 Markov Random Fields and Gibbs Random Fields 8.1.1 Markov Random Fields 8.1.2 Gibbs Random Fields 8.1.3 MRF-GRF Equivalence 8.1.4 Simplified Form of MRF 8.1.5 Generation of Texture Patterns Using MRF 8.2 Posterior Energy for Image Classification 8.3 Parameter Estimation 8.3.1 Least Squares Fit Method 8.3.2 Results of Parameter Estimations 8.4 MAP-MRF Classification Algorithms 8.4.1 Iterated Conditional Modes 8.4.2 Simulated Annealing 8.4.3 Maximizer of Posterior Marginals 8.5 Experimental Results Chapter 9: Multisource Classification 9.1 Image Fusion 9.1.1 Image Fusion Methods 9.1.2 Assessment of Fused Image Quality in the Spectral Domain 9.1.3 Performance Overview of Fusion Methods 9.2 Multisource Classification Using the Stacked-Vector Method 9.3 The Extension of Bayesian Classification Theory 9.3.1 An Overview 9.3.1.1 Feature Extraction 9.3.1.2 Probability or Evidence Generation 9.3.1.3 Multisource Consensus 9.3.2 Bayesian Multisource Classification Mechanism 9.3.3 A Refined Multisource Bayesian Model 9.3.4 Multisource Classification Using the Markov Random Field 9.3.5 Assumption of Intersource Independence 9.4 Evidential Reasoning 9.4.1 Concept Development 9.4.2 Belief Function and Belief Interval 9.4.3 Evidence Combination 9.4.4 Decision Rules for Evidential Reasoning 9.5 Dealing with Source Reliability 9.5.1 Using Classification Accuracy 9.5.2 Use of Class Separability 9.5.3 Data Information Class Correspondence Matrix 9.5.4 The Genetic Algorithm 9.6 Experimental Results Bibliography Index

Automated melanoma recognition

Abstract: A system for the computerized analysis of images obtained from epiluminescence microscopy (ELM) has been developed to enhance the early recognition of malignant melanoma. As an initial step, the binary mask of the skin lesion is determined by several basic segmentation algorithms together with a fusion strategy. A set of features containing shape and radiometric features as well as local and global parameters is calculated to describe the malignancy of a lesion. Significant features are then selected from this set by application of statistical feature subset selection methods. The final kNN classification delivers a sensitivity of 87% with a specificity of 92%.

Support Vector Tracking

Abstract: Support Vector Tracking (SVT) integrates the Support Vector Machine (SVM) classifier into an optic-flow based tracker. Instead of minimizing an intensity difference function between successive frames, SVT maximizes the SVM classification score. To account for large motions between successive frames, we build pyramids from the support vectors and use a coarse-to-fine approach in the classification stage. We show results of using a homogeneous quadratic polynomial kernel-SVT for vehicle tracking in image sequences.

Automated segmentation of multiple sclerosis lesions by model outlier detection

Abstract: This paper presents a fully automated algorithm for segmentation of multiple sclerosis (MS) lesions from multispectral magnetic resonance (MR) images. The method performs intensity-based tissue classification using a stochastic model for normal brain images and simultaneously detects MS lesions as outliers that are not well explained by the model. It corrects for MR field inhomogeneities, estimates tissue-specific intensity models from the data itself, and incorporates contextual information in the classification using a Markov random field. The results of the automated method are compared with lesion delineations by human experts, showing a high total lesion load correlation. When the degree of spatial correspondence between segmentations is taken into account, considerable disagreement is found, both between expect segmentations, and between expert and automatic measurements.

Support vector machines for SAR automatic target recognition

Abstract: Algorithms that produce classifiers with large margins, such as support vector machines (SVMs), AdaBoost, etc, are receiving more and more attention in the literature. A real application of SVMs for synthetic aperture radar automatic target recognition (SAR/ATR) is presented and the result is compared with conventional classifiers. The SVMs are tested for classification both in closed and open sets (recognition). Experimental results showed that SVMs outperform conventional classifiers in target classification. Moreover, SVMs with the Gaussian kernels are able to form a local "bounded" decision region around each class that presents better rejection to confusers.

Multiscale image segmentation using wavelet-domain hidden Markov models

Abstract: We introduce a new image texture segmentation algorithm, HMTseg, based on wavelets and the hidden Markov tree (HMT) model. The HMT is a tree-structured probabilistic graph that captures the statistical properties of the coefficients of the wavelet transform. Since the HMT is particularly well suited to images containing singularities (edges and ridges), it provides a good classifier for distinguishing between textures. Utilizing the inherent tree structure of the wavelet HMT and its fast training and likelihood computation algorithms, we perform texture classification at a range of different scales. We then fuse these multiscale classifications using a Bayesian probabilistic graph to obtain reliable final segmentations. Since HMTseg works on the wavelet transform of the image, it can directly segment wavelet-compressed images without the need for decompression into the space domain. We demonstrate the performance of HMTseg with synthetic, aerial photo, and document image segmentations.

Unsupervised classification of multifrequency and fully polarimetric SAR images based on the H/A/Alpha-Wishart classifier

Abstract: Introduces a new classification scheme for dual frequency polarimetric SAR data sets. A (6/spl times/6) polarimetric coherency matrix is defined to simultaneously take into account the full polarimetric information from both images. This matrix is composed of the two coherency matrices and their cross-correlation. A decomposition theorem is applied to both images to obtain 64 initial clusters based on their scattering characteristics. The data sets are then classified by an iterative algorithm based on a complex Wishart density function of the 6/spl times/6 matrix. A class number reduction technique is then applied on the 64 resulting clusters to improve the efficiency of the interpretation and representation of each class. An alternative technique is also proposed which introduces the polarimetric cross-correlation information to refine the results of classification to a small number of clusters using the conditional probability of the cross-correlation matrix. These classification schemes are applied to full polarimetric P, L, and C-band SAR images of the Nezer Forest, France, acquired by the NASA/JPL AIRSAR sensor in 1989.

Application of data mining techniques for medical image classification

Abstract: Breast cancer represents the second leading cause of cancer deaths in women today and it is the most common type of cancer in women. This paper presents some experiments for tumour detection in digital mammography. We investigate the use of different data mining techniques, neural networks and association rule mining, for anomaly detection and classification. The results show that the two approaches performed well, obtaining a classification accuracy reaching over 70% percent for both techniques. Moreover, the experiments we conducted demonstrate the use and effectiveness of association rule mining in image categorization.

Rule-Based Classification Systems Using Classification and Regression Tree (CART) Analysis

Abstract: Incorporating ancillary data into image classification can increase classification accuracy and precision. Rule-based classification systems using expert systems or machine learning are a particularly useful means of incorporating ancillary data, but have been difficult to implement. We developed a means for creating a rule-based classification using classification and regression tree analysis (CART), a commonly available statistical method. The CART classification does not require expert knowledge, automatically selects useful spectral and ancillary data from data supplied by the analyst, and can be used with continuous and categorical ancillary data. We demonstrated the use of the CART classification at three increasingly detailed classification levels for a portion of the Greater Yellowstone Ecosystem. Overall accuracies ranged from 96 percent at level 1, to 79 percent at level 2, and 65 percent at level 3.

Super-resolution target identification from remotely sensed images using a Hopfield neural network

Abstract: Fuzzy classification techniques have been developed recently to estimate the class composition of image pixels, but their output provides no indication of how these classes are distributed spatially within the instantaneous field of view represented by the pixel. As such, while the accuracy of land cover target identification has been improved using fuzzy classification, it remains for robust techniques that provide better spatial representation of land cover to be developed. Such techniques could provide more accurate land cover metrics for determining social or environmental policy, for example. The use of a Hopfield neural network to map the spatial distribution of classes more reliably using prior information of pixel composition determined from fuzzy classification was investigated. An approach was adopted that used the output from a fuzzy classification to constrain a Hopfield neural network formulated as an energy minimization tool. The network converges to a minimum of an energy function, defined as a goal and several constraints. Extracting the spatial distribution of target class components within each pixel was, therefore, formulated as a constraint satisfaction problem with an optimal solution determined by the minimum of the energy function. This energy minimum represents a "best guess" map of the spatial distribution of class components in each pixel. The technique was applied to both synthetic and simulated Landsat TM imagery, and the resultant maps provided an accurate and improved representation of the land covers studied, with root mean square errors (RMSEs) for Landsat imagery of the order of 0.09 pixels in the new fine resolution image recorded.

Support vector machine learning for image retrieval

Abstract: A novel method of relevance feedback is presented based on support vector machine learning in the content-based image retrieval system. A SVM classifier can be learned from training data of relevance images and irrelevance images marked by users. Using the classifier, the system can retrieve more images relevant to the query in the database efficiently. Experiments were carried out on a large-size database of 9918 images. It shows that the interactive learning and retrieval process can find correct images increasingly. It also shows the generalization ability of SVM under the condition of limited training samples.

Best-bases feature extraction algorithms for classification of hyperspectral data

Abstract: Due to advances in sensor technology, it is now possible to acquire hyperspectral data simultaneously in hundreds of bands. Algorithms that both reduce the dimensionality of the data sets and handle highly correlated bands are required to exploit the information in these data sets effectively. the authors propose a set of best-bases feature extraction algorithms that are simple, fast, and highly effective for classification of hyperspectral data. These techniques intelligently combine subsets of adjacent bands into a smaller number of features. Both top-down and bottom-up algorithms are proposed. The top-down algorithm recursively partitions the bands into two (not necessarily equal) sets of bands and then replaces each final set of bands by its mean value. The bottom-up algorithm builds an agglomerative tree by merging highly correlated adjacent bands and projecting them onto their Fisher direction, yielding high discrimination among classes. Both these algorithms are used in a pairwise classifier framework where the original C-class problem is divided into a set of (/sub 2//sup C/) two-class problems. The new algorithms (1) find variable length bases localized in wavelength, (2) favor grouping highly correlated adjacent bands that, when merged either by taking their mean or Fisher linear projection, yield maximum discrimination, and (3) seek orthogonal bases for each of the (/sub 2//sup C/) two-class problems into which a C-class problem can be decomposed. Experiments on an AVIRIS data set for a 12-class problem show significant improvements in classification accuracies while using a much smaller number of features.

A shape- and texture-based enhanced Fisher classifier for face recognition

Abstract: This paper introduces a new face coding and recognition method, the enhanced Fisher classifier (EFC), which employs the enhanced Fisher linear discriminant model (EFM) on integrated shape and texture features. Shape encodes the feature geometry of a face while texture provides a normalized shape-free image. The dimensionalities of the shape and the texture spaces are first reduced using principal component analysis, constrained by the EFM for enhanced generalization. The corresponding reduced shape and texture features are then combined through a normalization procedure to form the integrated features that are processed by the EFM for face recognition. Experimental results, using 600 face images corresponding to 200 subjects of varying illumination and facial expressions, show that (1) the integrated shape and texture features carry the most discriminating information followed in order by textures, masked images, and shape images, and (2) the new coding and face recognition method, EFC, performs the best among the eigenfaces method using L/sub 1/ or L/sub 2/ distance measure, and the Mahalanobis distance classifiers using a common covariance matrix for all classes or a pooled within-class covariance matrix. In particular, EFC achieves 98.5% recognition accuracy using only 25 features.

Linear image coding for regression and classification using the tensor-rank principle

Abstract: Given a collection of images (matrices) representing a "class" of objects we present a method for extracting the commonalities of the image space directly from the matrix representations (rather than from the vectorized representation which one would normally do in a PCA approach, for example). The general idea is to consider the collection of matrices as a tensor and to look for an approximation of its tensor-rank. The tensor-rank approximation is designed such that the SVD decomposition emerges in the special case where all the input matrices are the repeatition of a single matrix. We evaluate the coding technique both in terms of regression, i.e., the efficiency of the technique for functional approximation, and classification. We find that for regression the tensor-rank coding, as a dimensionality reduction technique, significantly outperforms other techniques like PCA. As for classification, the tensor-rank coding is at is best when the number of training examples is very small.

Shadow identification and classification using invariant color models

Abstract: A novel approach to shadow detection is presented. The method is based on the use of invariant color models to identify and to classify shadows in digital images. The procedure is divided into two levels: first, shadow candidate regions are extracted; then, by using the invariant color features, shadow candidate pixels are classified as self shadow points or as cast shadow points. The use of invariant color features allows a low complexity of the classification stage. Experimental results show that the method succeeds in detecting and classifying shadows within the environmental constrains assumed as hypotheses, which are less restrictive than state-of-the-art methods with respect to illumination conditions and the scene's layout.

Unsupervised retraining of a maximum likelihood classifier for the analysis of multitemporal remote sensing images

Abstract: An unsupervised retraining technique for a maximum likelihood (ML) classifier is presented. The proposed technique allows the classifier's parameters, obtained by supervised learning on a specific image, to be updated in a totally unsupervised way on the basis of the distribution of a new image to be classified. This enables the classifier to provide a high accuracy for the new image even when the corresponding training set is not available.

Improvement of mammographic mass characterization using spiculation measures and morphological features

Abstract: We are developing new computer vision techniques for characterization of breast masses on mammograms. We had previously developed a characterization method based on texture features. The goal of the present work was to improve our characterization method by making use of morphological features. Toward this goal, we have developed a fully automated, three-stage segmentation method that includes clustering, active contour, and spiculation detection stages. After segmentation, morphological features describing the shape of the mass were extracted. Texture features were also extracted from a band of pixels surrounding the mass. Stepwise feature selection and linear discriminant analysis were employed in the morphological, texture, and combined feature spaces for classifier design. The classification accuracy was evaluated using the area Az under the receiver operating characteristic curve. A data set containing 249 films from 102 patients was used. When the leave-one-case-out method was applied to partition the data set into trainers and testers, the average test Az for the task of classifying the mass on a single mammographic view was 0.83 +/- 0.02, 0.84 +/- 0.02, and 0.87 +/- 0.02 in the morphological, texture, and combined feature spaces, respectively. The improvement obtained by supplementing texture features with morphological features in classification was statistically significant (p = 0.04). For classifying a mass as malignant or benign, we combined the leave-one-case-out discriminant scores from different views of a mass to obtain a summary score. In this task, the test Az value using the combined feature space was 0.91 +/- 0.02. Our results indicate that combining texture features with morphological features extracted from automatically segmented mass boundaries will be an effective approach for computer-aided characterization of mammographic masses.

An approach to the automatic design of multiple classifier systems

Abstract: Multiple classifier systems (MCSs) based on the combination of outputs of a set of different classifiers have been proposed in the field of pattern recognition as a method for the development of high performance classification systems. Previous work clearly showed that multiple classifier systems are effective only if the classifiers forming them are accurate and make different errors. Therefore, the fundamental need for methods aimed to design “accurate and diverse” classifiers is currently acknowledged. In this paper, an approach to the automatic design of multiple classifier systems is proposed. Given an initial large set of classifiers, our approach is aimed at selecting the subset made up of the most accurate and diverse classifiers. A proof of the optimality of the proposed design approach is given. Reported results on the classification of multisensor remote sensing images show that this approach allows the design of effective multiple classifier systems.

Optimal image scaling using pixel classification

Abstract: We introduce a new approach to optimal image scaling called resolution synthesis (RS). In RS, the pixel being interpolated is first classified in the context of a window of neighboring pixels; and then the corresponding high-resolution pixels are obtained by filtering with coefficients that depend upon the classification. RS is based on a stochastic model explicitly reflecting the fact that pixels falls into different classes such as edges of different orientation and smooth textures. We present a simple derivation to show that RS generates the minimum mean-squared error (MMSE) estimate of the high-resolution image, given the low-resolution image. The parameters that specify the stochastic model must be estimated beforehand in a training procedure that we have formulated as an instance of the well-known expectation-maximization (EM) algorithm. We demonstrate that the model parameters generated during the training may be used to obtain superior results even for input images that were not used during the training.

Computer-aided characterization of mammographic masses: accuracy of mass segmentation and its effects on characterization

Abstract: Mass segmentation is used as the first step in many computer-aided diagnosis (CAD) systems for classification of breast masses as malignant or benign. The goal of this paper was to study the accuracy of an automated mass segmentation method developed in our laboratory, and to investigate the effect of the segmentation stage on the overall classification accuracy. The automated segmentation method was quantitatively compared with manual segmentation by two expert radiologists (R1 and R2) using three similarity or distance measures on a data set of 100 masses. The area overlap measures between R1 and R2, the computer and R1, and the computer and R2 were 0.76/spl plusmn/0.13,0.74 /spl plusmn/0.11, and 0.74/spl plusmn/0.13, respectively. The interobserver difference in these measures between the two radiologists was compared with the corresponding differences between the computer and the radiologists. Using three similarity measures and data from two radiologists, a total of six statistical tests were performed. The difference between the computer and the radiologist segmentation was significantly larger than the interobserver variability in only one test. Two sets of texture, morphological, and spiculation features, one based on the computer segmentation, and the other based on radiologist segmentation, were extracted from a data set of 249 films from 102 patients. A classifier based on stepwise feature selection and linear discriminant analysis was trained and tested using the two feature sets. The leave-one-case-out method was used for data sampling. For case-based classification, the area A/sub z/ under the receiver operating characteristic (ROC) curve was 0.89 and 0.88 for the feature sets based on the radiologist segmentation and computer segmentation, respectively. The difference between the two ROC curves was not statistically significant.

Computerized image analysis: estimation of breast density on mammograms.

Abstract: An automated image analysis tool is being developed for the estimation of mammographic breast density. This tool may be useful for risk estimation or for monitoring breast density change in prevention or intervention programs. In this preliminary study, a data set of 4-view mammograms from 65 patients was used to evaluate our approach. Breast density analysis was performed on the digitized mammograms in three stages. First, the breast region was segmented from the surrounding background by an automated breast boundary-tracking algorithm. Second, an adaptive dynamic range compression technique was applied to the breast image to reduce the range of the gray level distribution in the low frequency background and to enhance the differences in the characteristic features of the gray level histogram for breasts of different densities. Third, rule-based classification was used to classify the breast images into four classes according to the characteristic features of their gray level histogram. For each image, a gray level threshold was automatically determined to segment the dense tissue from the breast region. The area of segmented dense tissue as a percentage of the breast area was then estimated. To evaluate the performance of the algorithm, the computer segmentation results were compared to manual segmentation with interactive thresholding by five radiologists. A "true" percent dense area for each mammogram was obtained by averaging the manually segmented areas of the radiologists. We found that the histograms of 6% (8 CC and 8 MLO views) of the breast regions were misclassified by the computer, resulting in poor segmentation of the dense region. For the images with correct classification, the correlation between the computer-estimated percent dense area and the "truth" was 0.94 and 0.91, respectively, for CC and MLO views, with a mean bias of less than 2%. The mean biases of the five radiologists' visual estimates for the same images ranged from 0.1% to 11%. The results demonstrate the feasibility of estimating mammographic breast density using computer vision techniques and its potential to improve the accuracy and reproducibility of breast density estimation in comparison with the subjective visual assessment by radiologists.

Unsupervised multiresolution segmentation for images with low depth of field

Abstract: Unsupervised segmentation of images with low depth of field (DOF) is highly useful in various applications. This paper describes a novel multiresolution image segmentation algorithm for low DOF images. The algorithm is designed to separate a sharply focused object-of-interest from other foreground or background objects. The algorithm is fully automatic in that all parameters are image independent. A multi-scale approach based on high frequency wavelet coefficients and their statistics is used to perform context-dependent classification of individual blocks of the image. Unlike other edge-based approaches, our algorithm does not rely on the process of connecting object boundaries. The algorithm has achieved high accuracy when tested on more than 100 low DOF images, many with inhomogeneous foreground or background distractions. Compared with he state of the art algorithms, this new algorithm provides better accuracy at higher speed.

MPEG VBR video traffic modeling and classification using fuzzy technique

Abstract: We present an approach for MPEG variable bit rate (VBR) video modeling and classification using fuzzy techniques. We demonstrate that a type-2 fuzzy membership function, i.e., a Gaussian MF with uncertain variance, is most appropriate to model the log-value of I/P/B frame sizes in MPEG VBR video. The fuzzy c-means (FCM) method is used to obtain the mean and standard deviation (std) of T/P/B frame sizes when the frame category is unknown. We propose to use type-2 fuzzy logic classifiers (FLCs) to classify video traffic using compressed data. Five fuzzy classifiers and a Bayesian classifier are designed for video traffic classification, and the fuzzy classifiers are compared against the Bayesian classifier. Simulation results show that a type-2 fuzzy classifier in which the input is modeled as a type-2 fuzzy set and antecedent membership functions are modeled as type-2 fuzzy sets performs the best of the five classifiers when the testing video product is not included in the training products and a steepest descent algorithm is used to tune its parameters.

Multiscale Bayesian segmentation using a trainable context model

Abstract: Multiscale Bayesian approaches have attracted increasing attention for use in image segmentation. Generally, these methods tend to offer improved segmentation accuracy with reduced computational burden. Existing Bayesian segmentation methods use simple models of context designed to encourage large uniformly classified regions. Consequently, these context models have a limited ability to capture the complex contextual dependencies that are important in applications such as document segmentation. We propose a multiscale Bayesian segmentation algorithm which can effectively model complex aspects of both local and global contextual behavior. The model uses a Markov chain in scale to model the class labels that form the segmentation, but augments this Markov chain structure by incorporating tree based classifiers to model the transition probabilities between adjacent scales. The tree based classifier models complex transition rules with only a moderate number of parameters. One advantage to our segmentation algorithm is that it can be trained for specific segmentation applications by simply providing examples of images with their corresponding accurate segmentations. This makes the method flexible by allowing both the context and the image models to be adapted without modification of the basic algorithm. We illustrate the value of our approach with examples from document segmentation in which test, picture and background classes must be separated.

A fuzzy logic classification scheme for selecting and blending satellite ocean color algorithms

Abstract: An approach for selecting and blending bio-optical algorithms is demonstrated using an ocean color satellite image of the northwest Atlantic shelf. This approach is based on a fuzzy logic classification scheme applied to the satellite-derived water-leaving radiance data, and it is used to select and blend class-specific algorithms. Local in situ bio-optical data were used to characterize optically-distinct water classes a priori and to parameterize algorithms for each class. Although the algorithms can be of any type (empirical or analytical), this demonstration involves class-specific semi-analytic algorithms, which are the inverse of a radiance model. The semi-analytic algorithms retrieve three variables related to the concentrations of optically active constituents. When applied to a satellite image, the fuzzy logic approach involves three steps. First, a membership function is computed for each pixel and each class. This membership function expresses the likelihood that the measured radiance belongs to a class, with a known reflectance distribution. Thus, for each pixel, class memberships are assigned to the predetermined classes on the basis of the derived membership functions. Second, three variables are retrieved from each of the class-specific algorithms for which the pixel has membership. Third, the class memberships are used to weight the class specific retrievals to obtain a final blended retrieval for each pixel. This approach allows for graded transitions between water types, and blends separately tuned algorithms for different water masses without suffering from the "patchwork quilt" effect associated with hard-classification schemes.

Unsupervised target detection in hyperspectral images using projection pursuit

Abstract: The authors present a projection pursuit (PP) approach to target detection. Unlike most of developed target detection algorithms that require statistical models such as linear mixture, the proposed PP is to project a high dimensional data set into a low dimensional data space while retaining desired information of interest. It utilizes a projection index to explore projections of interestingness. For target detection applications in hyperspectral imagery, an interesting structure of an image scene is the one caused by man-made targets in a large unknown background. Such targets can be viewed as anomalies in an image scene due to the fact that their size is relatively small compared to their background surroundings. As a result, detecting small targets in an unknown image scene is reduced to finding the outliers of background distributions. It is known that "skewness," is defined by normalized third moment of the sample distribution, measures the asymmetry of the distribution and "kurtosis" is defined by normalized fourth moment of the sample distribution measures the flatness of the distribution. They both are susceptible to outliers. So, using skewness and kurtosis as a base to design a projection index may be effective for target detection. In order to find an optimal projection index, an evolutionary algorithm is also developed to avoid trapping local optima. The hyperspectral image experiments show that the proposed PP method provides an effective means for target detection.