International Journal of Pattern Recognition and Artificial Intelligence

Sign language recognition (SLR) has got wide applicability. SLR system is considered to be a challenging one. This paper presents empirical analysis of different mathematical models for Pakistan SLR (PSLR). The proposed method is using the parameterization of sign signature. Each sign is represented with a mathematical function and then coefficients of these functions are used as the feature vector. This approach is based on exhaustive experimentation and analysis for getting the best suitable mathematical representation for each sign. This extensive empirical analysis, results in a very small feature vector and hence to a very efficient system. The robust proposed method has got general applicability as it just need a new training set and it can work equally good for any other dataset. Sign set used is quite complex in the sense that intersign similarity distance is very small but even then proposed methodology has given quite promising results.

Driver fatigue and inattention have long been recognized as one of the main contributing factors in traffic accidents. Therefore, the development of intelligent driver assistance systems, which provides automatic monitoring of driver's vigilance, is an urgent and challenging task. This paper presents a novel system for video-based driving behavior recognition. The fundamental idea is to monitor driver's hand movements and to use these as predictors for safe/unsafe driving behavior. In comparison to previous work, the proposed method utilizes hierarchical classification and treats driving behavior in terms of a spatio-temporal reference framework as opposed to a static image. The approach was verified using the Southeast University Driving-Posture Dataset, a dataset comprised of video clips covering aspects of driving such as: normal driving, responding to a cell phone call, eating and smoking. After pre-processing for illumination variations and motion sequence segmentation, eight classes of behavior were identified. The overall prediction accuracy obtained using the proposed approach was $89.62\%$ when using a hierarchical classification approach. The proposed approach was able to clearly identify two dangerous driving behaviors, Responding to a cellphone call and Eating, with recognition rates of 92.39% and 92.29% respectively.

Day-ahead prediction of wind speed is a basic and key problem of large-scale wind power penetration. Many current techniques fail to satisfy practical engineering requirements because of wind speed's strong nonlinear features, influenced by many complex factors, and the general model's inability to automatically learn features. It is well recognized that wind speed varies in different patterns. In this paper, we propose a deep feature learning (DFL) approach to wind speed forecasting because of its advantages at both multi-layer feature extraction and unsupervised learning. A deep belief network (DBN) model for regression with an architecture of 144 input and 144 output nodes was constructed using a restricted Boltzmann machine (RBM). Day-ahead prediction experiments were then carried out. By comparing the experimental results, it was found that the prediction errors with respect to both size and stability of a DBN model with only three hidden layers were less than those of the other three typical approaches including support vector regression (SVR), single hidden layer neural networks (SHL-NN), and neural networks with three hidden layers (THL-NN). In addition, the DBN model can learn and obtain complex features of wind speed through its strong nonlinear mapping ability, which effectively improves its prediction precision. In addition, prediction errors are minimized when the number of DBN model's hidden layers reaches a threshold value. Above this number, it is not possible to improve the prediction accuracy by further increasing the number of hidden layers. Thus, the DBN method has a high practical value for wind speed prediction.

Image watermarking intends to hide secret data for the purposes of copyright protection, image authentication, data privacy, and broadcast monitoring. The ultimate goal is to achieve highest embedding capacity and lowest image distortion. In this paper, we present an intelligent watermarking scheme which can automatically analyze an image content to extract significant regions of interest (ROIs). A ROI is an area involving crucial information, and will be kept intact. The remaining regions of non-interest (RONIs) are collated for embedding watermarks, and will be ranked according to their entropy fuzzy memberships into a degree of importance. They are embedded by different amounts of bits to achieve optimal watermarking. The watermark is compressed and embedded in the frequency domain of the image. Experimental results show that the proposed technique has accomplished high capacity, high robustness, and high PSNR (peak signal-to-noise ratio) watermarking.

In real-time rendering transparency is an important multi-fragment effect to visualize the structure of three-dimensional models. The per-pixel transmittance implicitly describes how the light is attenuated by traveling through several transparent fragments. We present a hybrid approach to fit the transmittance using the Heaviside step function and the trigonometric function. The k fragments with the largest contribution are exactly composited and the remaining ones are accurately compressed in an unified formulation. With a single geometry pass, fragments are sorted into a fixed-size array and overflowing ones are expanded by a truncated Fourier series. Then the transmittance is reconstructed on the fly to modulate the surface color in another geometry pass. Our approach favors high scene complexity but operates in bounded memory without losing noticeable high-frequency detail. We demonstrate that it is able to closely match the image quality at competitive frame rate, comparing to a realtime A-buffer implementation and other approximate transparency techniques.

In this paper, we propose a novel method for cross-view action recognition via multiple continuous virtual paths which connect the source view and the target view. Each point on one virtual path is a virtual view which is obtained by a linear transformation of an action descriptor. All the virtual views are concatenated into an infinite-dimensional feature to characterize continuous changes from the source to the target view. To utilize these infinite-dimensional features directly, we propose a virtual view kernel (VVK) to compute the similarity between two infinite-dimensional features, which can be readily used to construct any kernelized classifiers. In addition, a constraint term is introduced to fully utilize the information contained in the unlabeled samples which are easier to obtain from the target view. The rationality behind the constraint is that any action video belongs to only one class. To further explore complementary visual information, we utilize multiple continuous virtual paths. The original source and target views are projected to different auxiliary source and target views using the random projection technique. Then we fuse all the VVKs generated from all pairs of auxiliary views. Our method is verified on the IXMAS and MuHAVi datasets, and the experimental results demonstrate that our method achieves better performance than the state-of-the-art methods.

In biomedical documents/publications, medical images tend to be complex by nature and often contain several regions that are annotated using arrows. In this context, an automated arrowhead detection is a critical precursor to region-of-interest (ROI) labeling and image content analysis. To detect arrowheads, in this paper, images are first binarized using fuzzy binarization technique to segment a set of candidates based on connected component (CC) principle. To select arrow candidates, we use convexity defect-based filtering, which is followed by template matching via dynamic time warping (DTW). The DTW similarity score confirms the presence of arrows in the image. Our test results on biomedical images from imageCLEF 2010 collection shows the interest of the technique, and can be compared with previously reported state-of-the-art results.

An optimization strategy of sound quality was proposed to improve sound quality inside a vehicle. The frequency of vehicle interior noise was divided into sub-bands by equivalent rectangular bandwidth (ERB) in the frequency domain, and the intensity of sub-band noise located in the mid-low frequency can be modified by using the measure of active noise control (ANC) in order to analyze its influence on the sound quality of original noise. Orthogonal experiments of vehicle interior noise were conducted according to experimental design. A model of sound quality objective evaluation (SQOE) was proposed as an experimental index. The role of sub-band on the model of SQOE was analyzed at different speeds so as to achieve the optimum level, optimum combination of sound quality and main factors. The results, obtained by implementing the optimization strategy of sound quality, were verified by means of tests. By validation, it showed that optimum sound quality can be achieved by changing the intensity of sub-band noise.

When used for optimizing complex functions, harmony search (HS) and shuffled frog leaping algorithm (SFLA) algorithm tend to easily get trapped into local optima and result in low convergence precision. To overcome such shortcomings, a hybrid mechanism of selective search by combining HS algorithm and SFLA algorithm is as well proposed. An HS-SFLA algorithm is designed by taking the advantages of HS and SFLA algorithms. The hybrid algorithm of HS-SFLA is adopted for dealing with complex function optimization problems, the experimental results show that HS-SFLA outperforms other state-of-the-art intelligence algorithms significantly in terms of global search ability, convergence speed and robustness on 80% of the benchmark functions tested. The HS-SFLA algorithm could directly be applied to all kinds of continuous optimization problems in the real world.

Intrusion detection is a kind of security mechanism which is used to detect attacks and intrusion behaviors. Due to the low accuracy and the high false positive rate of the existing clonal selection algorithms applied to intrusion detection, in this paper, we proposed a feature selection method for improved clonal algorithm. The improved method detects the intrusion behavior by selecting the best individual overall and clones them. Experimental results show that the feature selection algorithm is better than the traditional feature selection algorithm on the different classifiers, and it is shown that the final detection results are better than traditional clonal algorithm with 99.6% accuracy and 0.1% false positive rate.

This paper studies the group consensus problem in networks of multi-agent systems, in which the nonlinear dynamics of all agents are unknown and nonidentical. We assume that the unknown dynamics are linearly parameterized. Adaptive control method is adopted in the algorithm design. A novel algorithm is proposed for the multi-agent systems to reach group consensus via pinning control strategies, which only rely on the relative position information between neighboring agents. The stability and parameter convergence analysis are done based on Lyapunov theory, Barbalat's lemma, adaptive control theory and algebraic graph theory. Finally, a simulation example is given to validate our theoretical results.