Earticle

Reconstruction of missing parts or scratches of digital images is an important field used extensively in artwork restoration. This restoration can be done by using two approaches, image inpainting and texture synthesis. There are many techniques for the two pervious approaches that can carry out the process optimally and accurately. In this paper the advantages and disadvantages of most algorithms of the image inpainting approach are discussed. The modification to Oliveira inpainting model is introduced. This modification produces fast and good quality with one iteration without blur and removes large object with symmetric background.

The Kohonen self organizing map is an excellent tool in exploratory phase of data mining and pattern recognition. The SOM is a popular tool that maps high dimensional space into a small number of dimensions by placing similar elements close together, forming clusters. Recently researchers found that to capture the uncertainty involved in cluster analysis, it is not necessary to have crisp boundaries in some clustering operations. In this paper to overcome the uncertainty, a two-level clustering algorithm based on SOM which employs the rough set theory is proposed. The two-level stage Rough SOM (first using SOM to produce the prototypes that are then clustered in the second stage) is found to perform well and more accurate compared with the proposed crisp clustering method (Incremental SOM) and reduces the errors.

Multiple sequence alignment (MSA) is an important tool in biological analysis. However, it is difficult to solve this class of problems, due to their exponential complexity. This paper presents an algorithm combining the genetic algorithm and a self-organizing neural network for solution to MSA. This approach demonstrates improved performance in long DNA and RNA data sets exhibiting small similarity.

A method is proposed for learning to classify spam and nonspam emails. It combines the strategy of the Best Stepwise Feature Selection with a classifier of Euclidean nearest-neighbor. Each text email is first transformed into a vector of D-dimensional Euclidean space. Emails were divided into training and test sets in the manner of 10-fold crossvalidation. Three experiments were performed, and their elapsed CPU times and accuracies reported. The proposed spam detection learner was found to be extremely fast in recognition and with good error rates. It could be used as a baseline learning agent, in terms of CPU time and accuracy, against which other learning agents can be measured.

In this paper, we propose a 2D feature tracking method that is stable to long video sequences. To improve the stability of long track-ing, we use trajectory information about 2D features. We predict the expected feature states and compute a rough estimate of the feature lo-cation on the current image frame using the history of previous feature states up to the current frame. A search window is positioned at the esti-mated location and similarity measures are computed within the search window. Once the feature position is determined from the similarity mea-sures, the current feature states are appended to the history bu®er. The outlier rejection stage is also introduced to reduce false matches. Ex-perimental results from real video sequences showed that the proposed method stably tracks point features for long frame sequences.

We developed a ubiquitous healthcare system consisted of a physiological signal devices, a mobile system, a device provider system, a healthcare service provider system, a physician system, and a healthcare personal system. In this system, wireless body area network (WBAN) such as ZigBee is used to communicate between physiological signal devices and the mobile system. WBAN device needs a specific function for ubiquitous healthcare application. We propose a scanning algorithm, dynamic discovery and installation, reliable data transmission, device access control, and a healthcare profile for ubiquitous healthcare system.

The main purpose of communication is to transfer information from one corner to another of the world. The information is basically stored in forms of documents or files created on the basis of requirements. So, the randomness of creation and storage makes them unstructured in nature. As a consequence, data retrieval and modification become hard nut to crack. The data, that is required frequently, should maintain certain pattern. Otherwise, problems like retrieving erroneous data or anomalies in modification or time consumption in retrieving process may hike. As every problem has its own solution, these unstructured documents have also given the solution named unstructured document categorization. That means, the collected unstructured documents will be categorized based on some given constraints. This paper is a review which deals with different techniques like text and data mining, genetic algorithm, lexical chaining, binarization method to reach the fulfillment of desired unstructured document categorization appeared in the literature.

Skeletonization has been a part of morphological image processing for a wide variety of applications. Skeletonization algorithms have played an important role in the preprocessing phase of OCR systems. Many algorithms for vectorization by skeletonization have been devised and applied to a great variety of pictures and drawings for data compression, pattern recognition and raster-to-vector conversion. The vectorization algorithms often used in pattern recognition tasks also require one-pixel-wide lines as input. But parallel skeletonization algorithms which generate one-pixel-wide skeletons can have difficulty in preserving the connectivity of an image or generate spurious branches. In this paper an alternative parallel skeletonization algorithm has been developed and implemented. This algorithm is better than already existing algorithms in terms of connectivity and spurious branches. A few most common skeletonization algorithms have been implemented and evaluated on the basis of performance parameters and compared with newly developed algorithm.