International Journal of Pattern Recognition and Artificial Intelligence

We investigate the implementation of image categorization algorithms with a reject option, as a mean to enhance the system reliability and to attain a higher classification accuracy. A reject option is desired in many image-classification applications for which the system should abstain from making decisions on the most uncertain images. Based on the random subspace (RS) ensemble learning model, a highly reliable image classification scheme is proposed by applying RS support vector machine (SVM) ensemble. Being different to previous classifier ensembles which focus on increasing classification accuracy exclusively, the objective of the proposed SVM ensemble is to provide classification confidence and implement reject option to accommodate the situations where no decision should be made. The ensemble is created with four different feature descriptions, including local binary pattern (LBP), pyramid histogram of oriented gradient (PHOG), Gabor filtering and curvelet transform. The consensus degree from the ensemble's voting conforms to the confidence measure and the rejection option is accomplished accordingly when the confidence falls below a threshold. The reliable recognition scheme is empirically evaluated on three image categorization benchmark databases, including the face database created by Aleix Martinez and Robert Benavente (AR faces), a subset of Caltech-101 images for object classification, and 15 natural scene categories, all of which yielded consistently high reliable results, thus demonstrating the effectiveness of the proposed approach. For example, a 99.9% accuracy was obtained with a rejection rate of 2.5% for the AR faces, which exhibit promising potentials for real-world applications.

Dimension reduction methods including feature selection and feature extraction have played an important role in data mining and pattern recognition. In this study, we propose a novel unsupervised feature selection approach based on sparse representation theory, namely Sparsity Score (SS). Due to the sparse representation procedure, SS not only owns the global property of Variance Score (VS) and the local property of Laplacian Score (LS), but also possesses the discriminating nature. Experimental results, based on three well-known face datasets (Yale, ORL and CMU PIE), reveal that SS performs well in the evaluation of the feature significance, and it significantly outperforms VS and LS.

Breast cancer is one of the leading causes of death from cancer in Taiwan. In this paper, we propose a feature-based scheme composed of preprocessing, feature extraction and a fuzzy classifier for suspicious region detection and identification. In the preprocessing stage, we first extract regions of interest and then coarsely determine suspicious regions via candidate screening. Some features are extracted based on intra-slice, texture and inter-slice analysis techniques for suspicious region identification. Intra-slice analysis evaluates the intensity and size of suspicious regions. To find a precise region, we propose a region growing algorithm based on ellipse-based approximation. In texture analysis, some texture cues are extracted from spatial and wavelet domains and integrated as a combined texture feature by using a neural network. Inter-slice analysis is based on the continuity characteristic and consistency of a suspicious region's size; the objective is to verify the static behavior of suspicious regions. Several magnetic resonance imaging (MRI) cases are utilized to evaluate the performance of the proposed scheme. Experimental results demonstrate that our scheme can not only extract regions of interest but also identify tumors well from magnetic resonance images.

Multi-view learning aims to effectively learn from data represented by multiple independent sets of attributes, where each set is taken as one view of the original data. In real-world application, each view should be acquired in unequal cost. Taking web-page classification for example, it is cheaper to get the words on itself (view one) than to get the words contained in anchor texts of inbound hyper-links (view two). However, almost all the existing multi-view learning does not consider the cost of acquiring the views or the cost of evaluating them. In this paper, we support that different views should adopt different representations and lead to different acquisition cost. Thus we develop a new view-dependent cost different from the existing both class-dependent cost and example-dependent cost. To this end, we generalize the framework of multi-view learning with the cost-sensitive technique and further propose a Cost-sensitive Multi-View Learning Machine named CMVLM for short. In implementation, we take into account and measure both the acquisition cost and the discriminant scatter of each view. Then through eliminating the useless views with a predefined threshold, we use the reserved views to train the final classifier. The experimental results on a broad range of data sets including the benchmark UCI, image, and bioinformatics data sets validate that the proposed algorithm can effectively reduce the total cost and have a competitive even better classification performance. The contributions of this paper are that: (1) first proposing a view-dependent cost; (2) establishing a cost-sensitive multi-view learning framework; (3) developing a wrapper technique that is universal to most multiple kernel based classifier.

This paper proposes a combination weighting algorithm using relative entropy for document clustering. Combination weighting is widely used in multiple attribute decision making (MADM) problem. However, there exist two difficulties to hinder the applications of combination weighting on document clustering. First, combination weighting is based on the integration of subjective weighting and objective weighting. However, there are so many attributes in documents that the subjective weights which rely on manual annotation by experts are impracticable. Secondly, a document data object might contain hundreds or even thousands of features. It is an extremely time-consuming task to calculate the combination weights. To address the issues, we suggest to simplify the combination weighting by not distinguishing subjective weight and objective weight. Meanwhile, we choose relative entropy method to reduce running time. In our algorithm, we obtain a combination weight set with 14 combination forms. The experiments on real document data show that both on the AC/PR/RE measures and the mutual information (MI) measure, the proposed CWRE-sIB algorithm is superior to the original sequential information bottleneck (sIB) algorithm and a series of weighting-sIB algorithms, which are built by applying a single weighting scheme to the original sIB algorithm.

The aim of this paper is to give an overview of state-of-the-art human detection methods, classify them into reasonable categories and listing new trends in this field. Human detection in static images or video sequences is a challenging area of research with its main application in surveillance, defence, intelligent vehicles and robotics. In this survey paper, we classify the human detection methods based upon the mode of data acquisition and features used for detection of human. The review of different algorithms reported in last few years for human detection is tabulated. Also, we reviewed standard dataset for evaluation of human detection and studied the statistics about these dataset such as number of frames, challenging environments conditions, ground truth structure, camera details, etc. Different performance measures for human detection algorithms both qualitative and quantitative are listed out. Our survey gauges the gap between the present research and future requirements.

Multifractal theory has been widely used in different kinds of fields. In this paper, methods were proposed to extract two kinds of multifractal descriptors of gray series and two-dimensional surfaces for gray image based on the multifractal detrended fluctuation analysis. The proposed multifractal parameters can be well described by texture feature through the test of some textures. Three aspects of experiments have been conducted to verify the robustness of the proposed parameters, which include noise immunity, degree of image blurring and compression ratio. Comparisons were conducted between the proposed parameters and other kinds of texture feature parameters calculated by the standard multifractal analysis, the method of differential box counting and the methods of gray level co-occurrence matrix. Results demonstrate that the proposed exponents of H(2) and h(2) have great noise immunity and are robust to image compression and blurring.

This paper presents a new local region-based and local gradient-based active contour model in a variational level set formulation for image segmentation. The model consists of three parts: the local region term, the local gradient term and the regularization term. The local region term is insensitive to noise, while the local gradient term has better capability of localization than the local region. The energy minimization is achieved by the level set evolution in an iterative strategy. In each iteration, the local intensity and local gradient are updated and fed into the level set evolution. Comparative experiments show that our model achieves the better performance on the ultrasound images with noise and low signal-to-noise ratio than the local binary fitting (LBF) energy model.

In this paper, we consider the problem of robust face recognition using color information. In this context, sparse representation-based algorithms are the state-of-the-art solutions for gray facial images. We will integrate the existing sparse representation-based algorithms with color information and this integration can improve the previous performances significantly. Furthermore, we propose a new performance metric, namely the discriminativeness (DIS) to describe the recognition effectiveness for sparse representation algorithms. We find out that the richer information in color space can be used to increase the DIS, i.e. enhancing the robustness in face recognition. Extensive experiments have been conducted under different conditions, including various feature extractors, random pixel corruptions and occlusions on AR and GT databases, to demonstrate the advantages of using color information in robust face recognition. Detailed analysis is also included for each experiment to explain why and how color improve the robustness of different sparse representation-based methods.

Handwritten signatures are a common behavioral biometric. They are widely accepted for identification purposes, such as approbating and authenticating legal documents and financial contracts. The main challenge of signature verification is the high dimensionality of the signature features dataset that makes the corroboration procedure computationally costly. In this paper, we reduced the dimension of the input data with almost no loss of information. To this end, wavelet transform and fusion techniques were used to propose a new set of features. In addition, we introduced an effective feature selection technique, which was based on applying a filter box to find the most informative parts of the data and eliminate redundancies. These methods improved operating speeds and reduced memory usage, as shown by our empirical studies using the Signature Verification Competition 2004 (SVC04) database. We obtained a competitive Equal Error Ratio (EER) of 2.5%, with considerably fewer features. These results suggest that the proposed package is comparable with the state-of-the-art methods while using a significantly smaller number of features.