Boonwat Attachoo

Feature subset selection wrapper based on mutual information and rough sets

In this paper, we introduced a novel feature selection method based on the hybrid model (filter-wrapper). We developed a feature selection method using the mutual information criterion without requiring a user-defined parameter for the selection of the candidate feature set. Subsequently, to reduce the computational cost and avoid encountering to local maxima of wrapper search, a wrapper approach searches in the space of a superreduct which is selected from the candidate feature set. Finally, the wrapper approach determines to select a proper feature set which better suits the learning algorithm. The efficiency and effectiveness of our technique is demonstrated through extensive comparison with other representative methods. Our approach shows an excellent performance, not only high classification accuracy, but also with respect to the number of features selected.

A new approach for colored satellite image enhancement

A unique method of image filtering has been developed that enhances the detail and sharpens the edges of colored satellite images. Histogram equalization coupled with a two stage data filtering process that applies convolution with laplacian and sharpening with laplacaian through the 3 color bands that produce the colored satellite images has yielded sharper clearer images. The initial enhancement using histogram equalization was followed by the first stage of a filtering process convolution with laplacian which highlighted the edges of the image. The application of the second stage filtering sharpening with laplacian yielded enhanced color reproduction and a more accurate depiction of information at sea and land levels than was available in the original image. An analysis of the statistical index and signal to noise ratio of the true color and false color of histogram equalization, convolution with laplacian and sharpening images showed the image. An analysis of the false color of histogram equalization, convolution with laplacian and sharpening images showed the image to be superior, in this study the multi spectral content in the satellite image was transformed into a composite colour image, and then convoluted using the laplacian technique. This study placed an emphasis on improving the detail and edge clarity of satellite images.

Estimating Optimal Feature Subsets Using Mutual Information Feature Selector and Rough Sets

Mutual Information (MI) is a good selector of relevance between input and output feature and have been used as a measure for ranking features in several feature selection methods. Theses methods cannot estimate optimal feature subsets by themselves, but depend on user defined performance. In this paper, we propose estimation of optimal feature subsets by using rough sets to determine candidate feature subset which receives from MI feature selector. The experiment shows that we can correct nonlinear problems and problems in situation of two or more combined features are dominant features, maintain an improve classification accuracy.

Content-based image retrieval system based on combined and weighted multi-features

This paper, we proposed a novel framework for combining and weighting all of three i.e. color, shape and texture features to achieve higher retrieval efficiency. The color feature is extracted by quantifying the YUV color space and the color attributes like the mean value, the standard deviation, and the image bitmap of YUV color space is represented. The texture features are obtained by the entropy based on the gray level cooccurrence matrix and the edge histogram descriptor of an image. The shape feature descriptor is derived from Fourier descriptors (FDs) and the FDs derived from different signatures. When computing the similarity between the query image and target image in the database, normalization information distance is also used for adjusting distance values into the same level. And then the linear combination has used to combine the normalized distance of the color, shape and texture features to obtain the similarity as the indexing of image. Furthermore, an experimental results indicated, a weight variation to achieve higher retrieval efficiency and the proposed technique indeed outperforms other schemes in terms of the accuracy and efficiency.

A unified histogram and laplacian based for image sharpening

An effective method with enhancement procedures is proposed for image sharpening. Histogram equalization and edge detecting procedures are applied to original images. The mean value, standard deviation, and signal to noise ratio are defined as the statistical index which specifies the brightness, resolution, as well as the sharpness of the image. From the result of output images, the brightness and the contrast of the images were enhanced simultaneously with the sharpness and clearness of the mesh which outweigh the original image appearance. In addition, the advantages of this research indicate a suitable sharpening technique for the image category.

Growing rule-based induction system

Learning Classifier Systems (LCSs) are rule-based systems that have widely been used in data mining over the last few years. This paper employs UCS, a supervised learning classifier system, that was a version of LCSs for classification in data mining tasks. In this paper, we propose an adaptive framework of a rule-based competitive learning environment. In this framework, a growing neural gas (GNG) is used to adaptively cluster the data instances as they arrive. Each instance is then assigned to based classifier, the UCS responsible for the corresponding cluster. Through this mechanism, the complexity of a classification problem is decomposed adaptively into subproblems, each with a lower or equal complexity to the overall problem. Since each instance is exposed to a smaller population size than the single population approach, the throughput of the system increases. The experiments show that the proposed framework can decompose a problem adaptively into several subproblems. The accuracy rate of UCS in the distributed environment can also be better than the normal environment.