Vol. 38 · No. 6

In the fast evolution of storage systems, the newly emerged flash memory-based Solid State Drives (SSDs) are becoming an important part of the computer storage hierarchy. Amongst the several advantages of flash-based SSDs, high read performance, and low power consumption are of primary importance. Amongst its few disadvantages, its asymmetric I/O latencies for read, write and erase operations are the most crucial for overall performance. In this paper, we proposed two novel probabilistic adaptive algorithms that compute the future probability of reference based on recency, frequency, and periodicity of past page references. The page replacement is performed by considering the probability of reference of cached pages as well as asymmetric read-write-erase properties of flash devices. The experimental results show that our proposed method is successful in minimizing the performance overheads of flash-based systems as well as in maintaining the good hit ratio. The results also justify the utility of a genetic algorithm in maximizing the overall performance gains.

Most of the Visual Cryptography based image authentication schemes hide the share and authentication data into cover images by using an additional data hiding process. This process increases the computational cost of the schemes. Pixel expansion, meaningless shares and use of codebook are other challenges in these schemes. To overcome these issues, an authentication scheme is proposed in which no embedding into the cover images is performed and meaningful authentication shares are created using the watermark and cover images. This makes the scheme completely imperceptible. The watermark can be retrieved just by superimposing these authentication shares, thus reducing the computational complexity at receiver's side. Cellular Automata is used to construct the master share that provides self-construction ability to the shares. The meaningful authentication shares help in enhancing the security of the scheme while size invariance saves transmission and storage cost. The scheme possesses the ability of tamper detection and its localization. Experimental results demonstrate the improved security and quality of the generated shares of the proposed scheme as compared to existing schemes.

Sparse coding-based single image super-resolution has attracted much interest. In this paper, a super-resolution reconstruction algorithm based on sparse coding with multi-class dictionaries is put forward. We propose a novel method for image patch classification, using the phase congruency information. A sub-dictionary is learned from patches in each category. For a given image patch, the sub-dictionary that belongs to the same category is selected adaptively. Since the given patch has similar pattern with the selected sub-dictionary, it can be better represented. Finally, iterative back-projection is used to enforce global reconstruction constraint. Experiments demonstrate that our approach can produce comparable or even better super-resolution reconstruction results with some existing algorithms, in both subjective visual quality and numerical measures.

A new two-steps impulsive noise parallel Peer Group filter for color images using Compute Unified Device Architecture (CUDA) on a graphic card is proposed. It consists of two steps: impulsive noise detection, which uses a Fuzzy Metric as a distance criterion and a filtering step. For the needed ordering algorithm we are using the Marginal Median Filter with forgetful selection sort. Comparisons with other color filters for Graphics Processing Unit (GPU) architectures are presented, demonstrating that our proposal presents better performance in color preservation and noise suppression.

This paper presents a new method named AQueReBET, which automatically refines a query set by an information seeker searching on the web. A revelation of the intention of an information seeker who is running a search can bring a significant improvement to the search process and to browsing as well. It is practically impossible to acquire such intention by the explicit indication (feedback) due to the fact that web browsing takes place in real time. Therefore the intention must be determined in some other way. We hypothesize that it can be approximated by means of the implicit feedback preferably in the form of data from an eye tracker and mouse. We propose a method which automatically refines a seeker's search query, and thus we can offer documents with higher relevance, decrease the number of query reformulations and increase the seeker's satisfaction. The query refinement is based on an analysis of gaze data from an eye tracker and on groupization. In the proposed method, we calculate word-level importance based on term frequency, term uniqueness (tf-idf) and total fixation duration within the subdocument (word's snippet in search results).

The ease of creating digital content coupled with technological advancements allows institutions and organizations to further embrace distance learning. Teaching materials also receive attention, because it is difficult for the student to obtain adequate didactic material, being necessary a high effort and knowledge about the material and the repository. This work presents a framework that enables the automatic metadata generation for materials available in educational video repositories. Each module of the framework works autonomously and can be used in isolation, complemented by another technique or replaced by a more appropriate approach to the field of use, such as repositories with other types of media or other content.

Inference on Knowledge Bases (KBs) is an important way to construct more complete KBs and answer KB questions. Inference can be viewed as a process from evidence to conclusion following specific formulas. Traditional methods usually search on the KB to collect evidence, which cannot apply to large-scale KBs, because the running time of searching increases radically as the scale of KBs increases. What is worse, evidence cannot be found if one fact in it is missing, which may result in the failure of inference. To this end, we propose a fuzzy method of estimating evidence, which replaces searching by estimating the existence of evidence by constructing formula embeddings, and then we merge these estimations into a probabilistic model to infer conclusions. This method can apply to large-scale KBs, because estimating evidence is very fast and is irrelevant to the KB scale. Estimating evidence can also be viewed as fuzzy matching, so this method can handle the situation where facts are missing. We evaluate this method on the knowledge base completion task, and it achieves a better performance than state-of-the-art methods and has a shorter running time.

In this paper we introduce new selection method, 3-selection method. This method tries to generalize the most used selection methods in Genetic Algorithms (GA). Our new method involves both proportional and rank-based methods (order-based) and, moreover, it allows scaling of selection pressure with higher precision. This method is based on defining the shape of probability density distribution which is adjustable by parameters of our method. In addition, our method has one more attribute which adds randomness of selection.

A novel Chaotic Election Algorithm (CEA) is presented for numerical function optimization. CEA is a powerful enhancement of election algorithm. The election algorithm is a socio-politically inspired strategy that mimics the behavior of candidates and voters in presidential election process. In election algorithm, individuals are organized as electoral parties. Advertising campaign forms the basis of the algorithm in which individuals interact or compete with one other using three operators: positive advertisement, negative advertisement and coalition. Advertising campaign hopefully causes the individuals converge to the global optimum point in solution space. However, election algorithm suffers from a fundamental challenge: it gets stuck at local optima due to the inability of advertising campaign in searching solution space. CEA enhances the election algorithm through modifying party formation step, introducing chaotic positive advertisement and migration operator. By chaotic positive advertisement, CEA exploits the entire solution space, what increases the probability of obtaining global optimum point. By migration, CEA increases the diversity of the population and prevents early convergence of the individuals. The proposed CEA algorithm is tested on 28 well-known standard boundary-constrained test functions, and the results are verified by a comparative study with several well-known meta-heuristics. The results demonstrate that CEA is able to provide significant improvement over canonical election algorithm and other comparable algorithms.