International Journal of Pattern Recognition and Artificial Intelligence

Machine learning is among the most popular methods in designing unknown and variant malware detection algorithms. However, most of the existing methods take a single type of features to build binary classifiers. In practice, these methods have limited ability in depicting malware characteristics and the binary classification suffers from inadequate sampling of benign samples and extremely imbalanced training samples when detecting malware. In this paper, we present a malware detection Framework based on ENsemble One-Class Learning, namely FENOC. It uses hybrid features at different semantic layers to ensure a comprehensive insight of the program to be analyzed. We construct the malware detector by a novel learning algorithm called Cost-sensitive Twin One-class Classifier (CosTOC), which uses a pair of one-class classifiers to describe malware and benign programs respectively. CosTOC is more flexible and robust in comparison to conventional binary classifiers when training samples are extremely imbalanced or the benign programs are inadequately sampled. Finally, random subspace method and clustering-based ensemble method are developed to enhance the generalization ability of CosTOC. Experimental results show that FENOC gives a comparative detection rate and a lower false positive rate than many other binary classification algorithms, especially when the detector are trained with imbalanced data, or evaluated in terms of false positive rate.

Nowadays, it is very popular to use deep architectures in machine learning. Deep Belief Networks (DBNs) are deep architectures that use stack of Restricted Boltzmann Machines (RBMs) to create a powerful generative model using the training data. In this paper, we present an improvement in a common method that is usually used in training RBMs. The new method uses free energy as a criterion to obtain elite samples from generative model. We argue that these samples can compute gradient of log probability of training data more accurately. According to the results, an error rate of 0.99% was achieved on MNIST test set. This result shows that the proposed method outperforms the method presented in the first paper introducing DBN (1.25% error rate) and general classification methods such as SVM (1.4% error rate) and KNN (1.6% error rate). In another test using ISOLET dataset, letter classification error dropped to 3.59% compared to 5.59% error rate achieved in the papers using this dataset.

In this study, a semi-supervised learning algorithm for data classification and defect type recognition system for a 25 kV cross-linked polyethylene (XLPE) underground power cable joint is proposed. The proposed algorithm integrates the fuzzy-rough set and shared nearest neighbors (SNN) method for the assignment of labels to unlabeled data. As such, the data set is divided into two subsets: one is labeled and the other is unlabeled. The SNN is adopted for measuring the similarity between unlabeled data and labeled data subsets. Then, according to the levels of similarity, the fuzzy-rough set algorithm is adopted for assigning labels to the unlabeled data. A defect type recognition system XLPE cable classification problem is proposed in order to test the proposed algorithm. To demonstrate the performance of the proposed method, the proposed algorithm is applied to two well-known data sets. The experimental results show that the proposed algorithm can obtain outstanding levels of performance.

Arguably, the most difficult task in text classification is to choose an appropriate set of features that allows machine learning algorithms to provide accurate classification. Most state-of-the-art techniques for this task involve careful feature engineering and a pre-processing stage, which may be too expensive in the emerging context of massive collections of electronic texts. In this paper, we propose efficient methods for text classification based on information-theoretic dissimilarity measures, which are used to define dissimilarity-based representations. These methods dispense with any feature design or engineering, by mapping texts into a feature space using universal dissimilarity measures; in this space, classical classifiers (e.g. nearest neighbor or support vector machines) can then be used. The reported experimental evaluation of the proposed methods, on sentiment polarity analysis and authorship attribution problems, reveals that it approximates, sometimes even outperforms previous state-of-the-art techniques, despite being much simpler, in the sense that they do not require any text pre-processing or feature engineering.

In this paper, we address the problem of recognizing group activities of human objects based on their motion trajectory analysis. In order to resolve the complexity and ambiguity problems caused by a large number of human objects, we propose a Group Interaction Zone (GIZ) to detect meaningful groups in a scene to effectively handle noisy information. Two novel features, Group Interaction Energy (GIE) feature and Attraction and Repulsion Features, are proposed to better describe group activities within a GIZ. We demonstrate the performance of our method in two ways by (i) comparing the performance of the proposed method with the previous methods and (ii) analyzing the influence of the proposed features and GIZ-based meaningful group detection on group activity recognition using public datasets.

Microsoft Kinect, a low-cost motion sensing device, enables users to interact with computers or game consoles naturally through gestures and spoken commands without any other peripheral equipment. As such, it has commanded intense interests in research and development on the Kinect technology. In this paper, we present, a comprehensive survey on Kinect applications, and the latest research and development on motion recognition using data captured by the Kinect sensor. On the applications front, we review the applications of the Kinect technology in a variety of areas, including healthcare, education and performing arts, robotics, sign language recognition, retail services, workplace safety training, as well as 3D reconstructions. On the technology front, we provide an overview of the main features of both versions of the Kinect sensor together with the depth sensing technologies used, and review literatures on human motion recognition techniques used in Kinect applications. We provide a classification of motion recognition techniques to highlight the different approaches used in human motion recognition. Furthermore, we compile a list of publicly available Kinect datasets. These datasets are valuable resources for researchers to investigate better methods for human motion recognition and lower-level computer vision tasks such as segmentation, object detection and human pose estimation.

Clustering is a popular and effective method for image segmentation. However, existing cluster methods often suffer the following problems: (1) Need a huge space and a lot of computation when the input data are large. (2) Need to assign some parameters (e.g. number of clusters) in advance which will affect the clustering results greatly. To save the space and computation, reduce the sensitivity of the parameters, and improve the effectiveness and efficiency of the clustering algorithms, we construct a new clustering algorithm for image segmentation. The new algorithm consists of two phases: coarsening clustering and exact clustering. First, we use Affinity Propagation (AP) algorithm for coarsening. Specifically, in order to save the space and computational cost, we only compute the similarity between each point and its t nearest neighbors, and get a condensed similarity matrix (with only t columns, where t << N and N is the number of data points). Second, to further improve the efficiency and effectiveness of the proposed algorithm, the Self-tuning Spectral Clustering (SSC) is used to the resulted points (the representative points gotten in the first phase) to do the exact clustering. As a result, the proposed algorithm can quickly and precisely realize the clustering for texture image segmentation. The experimental results show that the proposed algorithm is more efficient than the compared algorithms FCM, K-means and SOM.

This paper presents a new technique for user identification and recognition based on the fusion of hand geometric features of both hands without any pose restrictions. All the features are extracted from normalized left and right hand images. Fusion is applied at feature and also at decision level. Two probability-based algorithms are proposed for classification. The first algorithm computes the maximum probability for nearest three neighbors. The second algorithm determines the maximum probability of the number of matched features with respect to a thresholding on distances. Based on these two highest probabilities initial decisions are made. The final decision is considered according to the highest probability as calculated by the Dempster–Shafer theory of evidence. Depending on the various combinations of the initial decisions, three schemes are experimented with 201 subjects for identification and verification. The correct identification rate is found to be 99.5%, and the false acceptance rate (FAR) of 0.625% has been found during verification.

Whilst facial recognition systems are vulnerable to different acquisition conditions, most notably lighting effects and pose variations, their particular level of sensitivity to facial aging effects is yet to be researched. The face recognition vendor test (FRVT) 2012's annual statement estimated deterioration in the performance of face recognition systems due to facial aging. There was about 5% degradation in the accuracies of the face recognition systems for each single year age difference between a test image and a probe image. Consequently, developing an age-invariant platform continues to be a significant requirement for building an effective facial recognition system. The main objective of this work is to address the challenge of facial aging which affects the performance of facial recognition systems. Accordingly, this work presents a geometrical model that is based on extracting a number of triangular facial features. The proposed model comprises a total of six triangular areas connecting and surrounding the main facial features (i.e. eyes, nose and mouth). Furthermore, a set of thirty mathematical relationships are developed and used for building a feature vector for each sample image. The areas and perimeters of the extracted triangular areas are calculated and used as inputs for the developed mathematical relationships. The performance of the system is evaluated over the publicly available face and gesture recognition research network (FG-NET) face aging database. The performance of the system is compared with that of some of the state-of-the-art face recognition methods and state-of-the-art age-invariant face recognition systems. Our proposed system yielded a good performance in term of classification accuracy of more than 94%.

Some regions (or blocks) and their affiliated features of face images are normally of more importance for face recognition. However, the variety of feature contributions, which exerts different saliency on recognition, is usually ignored. This paper proposes a new sparse facial feature description model based on salience evaluation of regions and features, which not only considers the contributions of different face regions, but also distinguishes that of different features in the same region. Specifically, the structured sparse learning scheme is employed as the salience evaluation method to encourage sparsity at both the group and individual levels for balancing regions and features. Therefore, the new facial feature description model is obtained by combining the salience evaluation method with region-based features. Experimental results show that the proposed model achieves better performance with much lower feature dimensionality.

The crucial problem that has to be solved when designing an effective brain–computer interface (BCI) is: how to reduce the huge space of features extracted from raw electroencephalography (EEG) signals. One of the strategies for feature selection that is often applied by BCI researchers is based on genetic algorithms (GAs). The two types of GAs that are most commonly used in BCI research are the classic algorithm and the Culling algorithm. This paper presents both algorithms and their application for selecting features crucial for the correct classification of EEG signals recorded during imagery movements of the left and right hand. The results returned by both algorithms are compared to those returned by an algorithm with aggressive mutation and an algorithm with melting individuals, both of which have been proposed by the author of this paper. While the aggressive mutation algorithm has been published previously, the melting individuals algorithm is presented here for the first time.