Showing papers in "Pattern Recognition Letters in 2006"

An introduction to ROC analysis

Abstract: Receiver operating characteristics (ROC) graphs are useful for organizing classifiers and visualizing their performance. ROC graphs are commonly used in medical decision making, and in recent years have been used increasingly in machine learning and data mining research. Although ROC graphs are apparently simple, there are some common misconceptions and pitfalls when using them in practice. The purpose of this article is to serve as an introduction to ROC graphs and as a guide for using them in research.

Random Forests for land cover classification

Abstract: Random Forests are considered for classification of multisource remote sensing and geographic data. Various ensemble classification methods have been proposed in recent years. These methods have been proven to improve classification accuracy considerably. The most widely used ensemble methods are boosting and bagging. Boosting is based on sample re-weighting but bagging uses bootstrapping. The Random Forest classifier uses bagging, or bootstrap aggregating, to form an ensemble of classification and regression tree (CART)-like classifiers. In addition, it searches only a random subset of the variables for a split at each CART node, in order to minimize the correlation between the classifiers in the ensemble. This method is not sensitive to noise or overtraining, as the resampling is not based on weighting. Furthermore, it is computationally much lighter than methods based on boosting and somewhat lighter than simple bagging. In the paper, the use of the Random Forest classifier for land cover classification is explored. We compare the accuracy of the Random Forest classifier to other better-known ensemble methods on multisource remote sensing and geographic data.

Efficient adaptive density estimation per image pixel for the task of background subtraction

Abstract: We analyze the computer vision task of pixel-level background subtraction. We present recursive equations that are used to constantly update the parameters of a Gaussian mixture model and to simultaneously select the appropriate number of components for each pixel. We also present a simple non-parametric adaptive density estimation method. The two methods are compared with each other and with some previously proposed algorithms.

Computer vision based method for real-time fire and flame detection

Abstract: This paper proposes a novel method to detect fire and/or flames in real-time by processing the video data generated by an ordinary camera monitoring a scene. In addition to ordinary motion and color clues, flame and fire flicker is detected by analyzing the video in the wavelet domain. Quasi-periodic behavior in flame boundaries is detected by performing temporal wavelet transform. Color variations in flame regions are detected by computing the spatial wavelet transform of moving fire-colored regions. Another clue used in the fire detection algorithm is the irregularity of the boundary of the fire-colored region. All of the above clues are combined to reach a final decision. Experimental results show that the proposed method is very successful in detecting fire and/or flames. In addition, it drastically reduces the false alarms issued to ordinary fire-colored moving objects as compared to the methods using only motion and color clues.

Automatic thresholding for defect detection

Abstract: Automatic thresholding has been widely used in the machine vision industry for automated visual inspection of defects. A commonly used thresholding technique, the Otsu method, provides satisfactory results for thresholding an image with a histogram of bimodal distribution. This method, however, fails if the histogram is unimodal or close to unimodal. For defect detection applications, defects can range from no defect to small or large defects, which means that the gray-level distributions range from unimodal to bimodal. For this paper, we revised the Otsu method for selecting optimal threshold values for both unimodal and bimodal distributions, and tested the performance of the revised method, the valley-emphasis method, on common defect detection applications.

Information-preserving hybrid data reduction based on fuzzy-rough techniques

Abstract: Data reduction plays an important role in machine learning and pattern recognition with a high-dimensional data. In real-world applications data usually exists with hybrid formats, and a unified data reducing technique for hybrid data is desirable. In this paper, an information measure is proposed to computing discernibility power of a crisp equivalence relation or a fuzzy one, which is the key concept in classical rough set model and fuzzy-rough set model. Based on the information measure, a general definition of significance of nominal, numeric and fuzzy attributes is presented. We redefine the independence of hybrid attribute subset, reduct, and relative reduct. Then two greedy reduction algorithms for unsupervised and supervised data dimensionality reduction based on the proposed information measure are constructed. Experiments show the reducts found by the proposed algorithms get a better performance compared with classical rough set approaches.

On-line trajectory clustering for anomalous events detection

Abstract: In this paper, we propose a trajectory clustering algorithm suited for video surveillance systems. Trajectories are clustered on-line, as the data are collected, and clusters are organized in a tree-like structure that, augmented with probability information, can be used to perform behaviour analysis, since it allows the identification of anomalous events.

Texture classification using Gabor wavelets based rotation invariant features

Abstract: Texture based image analysis techniques have been widely employed in the interpretation of earth cover images obtained using remote sensing techniques, seismic trace images, medical images and in query by content in large image data bases. The development in multi-resolution analysis such as wavelet transform leads to the development of adequate tools to characterize different scales of textures effectively. But, the wavelet transform lacks in its ability to decompose input image into multiple orientations and this limits their application to rotation invariant image analysis. This paper presents a new approach for rotation invariant texture classification using Gabor wavelets. Gabor wavelets are the mathematical model of visual cortical cells of mammalian brain and using this, an image can be decomposed into multiple scales and multiple orientations. The Gabor function has been recognized as a very useful tool in texture analysis, due to its optimal localization properties in both spatial and frequency domain and found widespread use in computer vision. Texture features are found by calculating the mean and variance of the Gabor filtered image. Rotation normalization is achieved by the circular shift of the feature elements, so that all images have the same dominant direction. The texture similarity measurement of the query image and the target image in the database is computed by minimum distance criterion.

Secret image sharing with smaller shadow images

Abstract: Secret image sharing is a technique for protecting images that involves the dispersion of the secret image into many shadow images. This endows the method with a higher tolerance against data corruption or loss than other image-protection mechanisms, such as encryption or steganography. In the method proposed in this study, the difference image of the secret image is encoded using Huffman coding scheme, and the arithmetic calculations of the sharing functions are evaluated in a power-of-two Galois Field GF(2^t). Experiment results show that each generated shadow image in the proposed method is about 40% smaller than that of the method in [Thien, C.C., Lin, J.C., 2002. Secret image sharing. Comput. Graphics 26 (1), 765-770], which improves its efficiency in storage, transmission, and data hiding.

Image segmentation by histogram thresholding using hierarchical cluster analysis

Abstract: This paper proposes a new method of image thresholding by using cluster organization from the histogram of an image. A new similarity measure proposed is based on inter-class variance of the clusters to be merged and the intra-class variance of the new merged cluster. Experiments on practical images illustrate the effectiveness of the new method.

Circle detection on images using genetic algorithms

Abstract: In this paper, we present a circle detection method based on genetic algorithms. Our genetic algorithm uses the encoding of three edge points as the chromosome of candidate circles (x,y,r) in the edge image of the scene. Fitness function evaluates if these candidate circles are really present in the edge image. Our encoding scheme reduces the search space by avoiding trying unfeasible individuals, this results in a fast circle detector. Our approach detects circles with sub-pixellic accuracy on synthetic images. Our method can also detect circles on natural images with sub-pixellic precision. Partially occluded circles can be located in both synthetic and natural images. Examples of the application of our method to the recognition of hand-drawn circles are also shown. Detection of several circles in a single image is also handled by our method.

Random subspace method for multivariate feature selection

Abstract: In a growing number of domains data captured encapsulates as many features as possible. This poses a challenge to classical pattern recognition techniques, since the number of samples often still is limited with respect to the number of features. Classical pattern recognition methods suffer from the small sample size, and robust classification techniques are needed. In order to reduce the dimensionality of the feature space, the selection of informative features becomes an essential step towards the classification task. The relevance of the features can be evaluated either individually (univariate approaches), or in a multivariate manner. Univariate approaches are simple and fast, therefore appealing. However, possible correlation and dependencies between the features are not considered. Therefore, multivariate search techniques may be helpful. Several limitations restrict the use of multivariate searches. First, they are prone to overtraining, especially in [email protected]?n (many features and few samples) settings. Secondly, they can be computationally too expensive when dealing with a large feature space. We introduce a new multivariate search technique, that is less sensitive to the noise in the data and computationally feasible as well. We compare our approach with several multivariate and univariate feature selection techniques, on an artificial dataset which provides us with ground truth information, and on a real dataset. The results show the importance of multivariate search techniques and the robustness and reliability of our novel multivariate feature selection method.

A local mean-based nonparametric classifier

Abstract: A considerable amount of effort has been devoted to design a classifier in practical situations. In this paper, a simple nonparametric classifier based on the local mean vectors is proposed. The proposed classifier is compared with the 1-NN, k-NN, Euclidean distance (ED), Parzen, and artificial neural network (ANN) classifiers in terms of the error rate on the unknown patterns, particularly in small training sample size situations. Experimental results show that the proposed classifier is promising even in practical situations.

Adaptive Hausdorff distances and dynamic clustering of symbolic interval data

Abstract: This paper presents a partitional dynamic clustering method for interval data based on adaptive Hausdorff distances. Dynamic clustering algorithms are iterative two-step relocation algorithms involving the construction of the clusters at each iteration and the identification of a suitable representation or prototype (means, axes, probability laws, groups of elements, etc.) for each cluster by locally optimizing an adequacy criterion that measures the fitting between the clusters and their corresponding representatives. In this paper, each pattern is represented by a vector of intervals. Adaptive Hausdorff distances are the measures used to compare two interval vectors. Adaptive distances at each iteration change for each cluster according to its intra-class structure. The advantage of these adaptive distances is that the clustering algorithm is able to recognize clusters of different shapes and sizes. To evaluate this method, experiments with real and synthetic interval data sets were performed. The evaluation is based on an external cluster validity index (corrected Rand index) in a framework of a Monte Carlo experiment with 100 replications. These experiments showed the usefulness of the proposed method.

Fault diagnosis of machines based on D-S evidence theory. Part 1: D-S evidence theory and its improvement

Abstract: In this paper, conventional D-S evidence theory is improved through the introduction of fuzzy membership function, importance index, and conflict factor in order to address the issues of evidence sufficiency, evidence importance, and conflicting evidences in the practical application of D-S evidence theory. New decision rules based on the improved D-S evidence theory are proposed. Examples are given to illustrate that the improved D-S evidence theory is better able to perform fault diagnosis through fusing multi-source information than conventional D-S evidence theory.

A region-based multi-sensor image fusion scheme using pulse-coupled neural network

Abstract: For most image fusion algorithms split relationship among pixels and treat them more or less independently, this paper proposes a region-based image fusion scheme using pulse-coupled neural network (PCNN), which combines aspects of feature and pixel-level fusion. The basic idea is to segment all different input images by PCNN and to use this segmentation to guide the fusion process. In order to determine PCNN parameters adaptively, this paper brings forward an adaptive segmentation algorithm based on a modified PCNN with the multi-thresholds determined by a novel water region area method. Experimental results demonstrate that the proposed fusion scheme has extensive application scope and it outperforms the multi-scale decomposition based fusion approaches, both in visual effect and objective evaluation criteria, particularly when there is movement in the objects or mis-registration of the source images.

A clustering-based method for unsupervised intrusion detections

Abstract: Detection of intrusion attacks is an important issue in network security. This paper considers the outlier factor of clusters for measuring the deviation degree of a cluster. A novel method is proposed to compute the cluster radius threshold. The data classification is performed by an improved nearest neighbor (INN) method. A powerful clustering-based method is presented for the unsupervised intrusion detection (CBUID). The time complexity of CBUID is linear with the size of dataset and the number of attributes. The experiments demonstrate that our method outperforms the existing methods in terms of accuracy and detecting unknown intrusions.

An efficient illumination normalization method for face recognition

Abstract: In this paper, an efficient representation method insensitive to varying illumination is proposed for human face recognition. Theoretical analysis based on the human face model and the illumination model shows that the effects of varying lighting on a human face image can be modeled by a sequence of multiplicative and additive noises. Instead of computing these noises, which is very difficult for real applications, we aim to reduce or even remove their effect. In our method, a local normalization technique is applied to an image, which can effectively and efficiently eliminate the effect of uneven illuminations while keeping the local statistical properties of the processed image the same as in the corresponding image under normal lighting condition. After processing, the image under varying illumination will have similar pixel values to the corresponding image that is under normal lighting condition. Then, the processed images are used for face recognition. The proposed algorithm has been evaluated based on the Yale database, the AR database, the PIE database, the YaleB database and the combined database by using different face recognition methods such as PCA, ICA and Gabor wavelets. Consistent and promising results were obtained, which show that our method can effectively eliminate the effect of uneven illumination and greatly improve the recognition results.

A watermarking-based image ownership and tampering authentication scheme

Abstract: Nowadays, image authentication schemes are widely applied to ownership protection and tampering detection of digital images. In this paper, we propose an image ownership and tampering authentication scheme based on watermarking techniques, which is employed to protect the rightful ownership and detect malicious manipulation over embedded images. In the scheme, the authentication data of an image will be inserted into adaptive least significant bits of the embedded pixels. And the number of the least significant bits will be determined according to the type of the corresponding block. The experimental results show that the quality of the embedded image is very high, and the positions of the tampered parts are located correctly.

Image thresholding using two-dimensional Tsallis-Havrda-Charvát entropy

Abstract: In this paper, we present a thresholding technique based on two-dimensional Tsallis-Havrda-Charvat entropy. The effectiveness of the proposed method is demonstrated by using examples from the real-world and synthetic images.

Selection of the optimal parameter value for the Isomap algorithm

Abstract: The isometric feature mapping (Isomap) method has demonstrated promising results in finding low-dimensional manifolds from data points in high-dimensional input space. Isomap has one free parameter (number of nearest neighbours K or neighbourhood radius @e), which has to be specified manually. In this paper we present a new method for selecting the optimal parameter value for Isomap automatically. Numerous experiments on synthetic and real data sets show the effectiveness of our method.

On minimum variance thresholding

Abstract: Variance-based thresholding methods could be biased from the threshold found by expert and the underlying mechanism responsible for this bias is explored in this paper. An analysis on the minimum class variance thresholding (MCVT) and the Otsu method, which minimizes the within-class variance, is carried out. It turns out that the bias for the Otsu method is due to differences in class variances or class probabilities and the resulting threshold is biased towards the component with larger class variance or larger class probability. The MCVT method is found to be similar to the minimum error thresholding.

Multi-class ROC analysis from a multi-objective optimisation perspective

Abstract: The receiver operating characteristic (ROC) has become a standard tool for the analysis and comparison of classifiers when the costs of misclassification are unknown. There has been relatively little work, however, examining ROC for more than two classes. Here we discuss and present an extension to the standard two-class ROC for multi-class problems. We define the ROC surface for the Q-class problem in terms of a multi-objective optimisation problem in which the goal is to simultaneously minimise the Q(Q-1) misclassification rates, when the misclassification costs and parameters governing the classifier's behaviour are unknown. We present an evolutionary algorithm to locate the Pareto front-the optimal trade-off surface between misclassifications of different types. The use of the Pareto optimal surface to compare classifiers is discussed and we present a straightforward multi-class analogue of the Gini coefficient. The performance of the evolutionary algorithm is illustrated on a synthetic three class problem, for both k-nearest neighbour and multi-layer perceptron classifiers.

Personal authentication using hand images

Abstract: This paper presents a new approach for personal authentication using hand images. The proposed method attempts to improve the performance of palmprint-based verification system by integrating hand geometry features. Unlike prior bimodal biometric systems, the users do not have to undergo the inconvenience of using two different sensors in our system since the palmprint and hand geometry images are acquired simultaneously using a single camera. The palmprint and handshape images are used to extract salient features and are then examined for their individual and combined verification performances. The image acquisition setup used here is inherently simple and it does not employ any special illumination nor does it use any alignment pegs to cause any inconvenience to the users. Our experiments on an image database of 100 users achieve promising results and suggest that the fusion of matching scores can achieve better performance than the fusion at representation.

Similarity measures on three kinds of fuzzy sets

Abstract: Intuitionistic fuzzy sets (IFSs) proposed by Atanassov, fuzzy rough sets (FRSs) proposed by Nanda and Majumdar, rough fuzzy sets (RFSs) proposed by Banerjee and Pal, have gained attention from researchers for their applications in various fields. Then similarity measures between three fuzzy sets were developed. In this paper, we first point out: IFSs, FRSs and RFSs are L-fuzzy sets with L being a special fuzzy lattice. At the same time, we suggest some rules which is considered when we give a similarity measure for measuring the degree of similarity between elements and between some fuzzy sets. After that, some existing measures of similarity are reviewed, some examples are applied to show that some existing similarity measures are not always effective in some cases. We propose some new similarity measures for measuring the degree of similarity between three fuzzy sets under an unifying form and between IFSs. Finally, we illustrate the problem in the context of colorectal cancer diagnosis by similarity measure between fuzzy rough sets. Therefore, the proposed similarity measures can provide a useful way for measuring three fuzzy sets more effectively.

An objective approach to cluster validation

Abstract: Cluster validation is a major issue in cluster analysis. Many existing validity indices do not perform well when clusters overlap or there is significant variation in their covariance structure. The contribution of this paper is twofold. First, we propose a new validity index for fuzzy clustering. Second, we present a new approach for the objective evaluation of validity indices and clustering algorithms. Our validity index makes use of the covariance structure of clusters, while the evaluation approach utilizes a new concept of overlap rate that gives a formal measure of the difficulty of distinguishing between overlapping clusters. We have carried out experimental studies using data sets containing clusters of different shapes and densities and various overlap rates, in order to show how validity indices behave when clusters become less and less separable. Finally, the effectiveness of the new validity index is also demonstrated on a number of real-life data sets.

A Bayes-spectral-entropy-based measure of camera focus using a discrete cosine transform

Abstract: In this paper we present a novel measure of camera focus based on the Bayes spectral entropy of an image spectrum. In order to estimate the degree of focus, the image is divided into non-overlapping sub-images of 8x8 pixels. Next, sharpness values are calculated separately for each sub-image and their mean is taken as a measure of the overall focus. The sub-image spectra are obtained by an 8x8 discrete cosine transform (DCT). Comparisons were made against four well-known measures that were chosen as reference, on images captured with a standard visible-light camera and a thermal camera. The proposed measure outperformed the reference measures by exhibiting a wider working range and a smaller failure rate. To assess its robustness to noise, additional tests were conducted with noisy images.

An adaptive real-time skin detector based on Hue thresholding: A comparison on two motion tracking methods

Abstract: Various applications like face and hand tracking and image retrieval have made skin detection an important area of research. However, currently available algorithms are based on static features of the skin colour, or require a significant amount of computation. On the other hand, skin detection algorithms are not robust to deal with real-world conditions, like background noise, change of intensity and lighting effects. This situation can be improved by using dynamic features of the skin colour in a sequence of images. This article proposes a skin detection algorithm based on adaptive Hue thresholding and its evaluation using two motion detection technique. The skin classifier is based on the Hue histogram of skin pixels, and adapts itself to the colour of the skin of the persons in the video sequence. This algorithm has demonstrated improvement in comparison to the static skin detection method.

Combining local features for robust nose location in 3D facial data

Abstract: Due to the wide use of human face images, it is significant to locate facial feature points. In this paper, we focus on 3D facial data and propose a novel method to solve a specific problem, i.e., locating the nose tip by one hierarchical filtering scheme combining local features. Based on the detected nose tip, we further estimate the nose ridge by a newly defined curve, the Included Angle Curve (IAC). The key features of our method are its automated implementation for detection, its ability to deal with noisy and incomplete input data, its invariance to rotation and translation, and its adaptability to different resolutions. The experimental results from different databases show the robustness and feasibility of the proposed method.

Automatic recognition of animal vocalizations using averaged MFCC and linear discriminant analysis

Abstract: In this paper we propose a method that uses the averaged Mel-frequency cepstral coefficients (MFCCs) and linear discriminant analysis (LDA) to automatically identify animals from their sounds. First, each syllable corresponding to a piece of vocalization is segmented. The averaged MFCCs over all frames in a syllable are calculated as the vocalization features. Linear discriminant analysis (LDA), which finds out a transformation matrix that minimizes the within-class distance and maximizes the between-class distance, is utilized to increase the classification accuracy while to reduce the dimensionality of the feature vectors. In our experiment, the average classification accuracy is 96.8% and 98.1% for 30 kinds of frog calls and 19 kinds of cricket calls, respectively.