• Journal of IKEEE
• /
• v.20 no.3
• /
• pp.209-219
• /
• 2016

SVM is a machine learning method used for image processing. It is well known for its high classification performance. We have to perform multiple MAC operations in order to use SVM for image classification. However, if the resolution of the target image or the number of classification cases increases, the execution time of SVM also increases, which makes it difficult to be performed in real-time applications. In this paper, we propose an hardware architecture which enables real-time applications using SVM classification. We used parallel architecture to simultaneously calculate MAC operations, and also designed the system for several feature extractors for compatibility. RBF kernel was used for hardware implemenation, and the exponent calculation formular included in the kernel was modified to enable fixed point modelling. Experimental results for the system, when implemented in Xilinx ZC-706 evaluation board, show that it can process 60.46 fps for $1360{\times}800$ resolution at 100MHz clock frequency.

• Journal of Internet Computing and Services
• /
• v.11 no.4
• /
• pp.41-49
• /
• 2010

Faloustsos et al.[1,2] showed successfully that the power-laws describe the heavy-tailed distributions of the AS level Internet topology properties such as node degree. This result allows us to represent the characteristics of AS-level Internet topology using some power-law exponents with elegant and simple. In this paper, we obtained the power-law exponents of the domestic AS-level Internet topology properties - the node degree, the number of pairs within hops, and eigenvalues of the graph - based on the valley-free BGP routing policy. We used the real data sets from UCLA IRL laboratory, and showed that these power-laws fit the real data pretty well resulting in correlation coefficient of 90.7%, 96.5%, and 97%, respectively. In particular, rounding the effective diameter to three, approximately 91% of the pairs of nodes are within this distance; that is, we might conclude our topology is pretty well organized.

• Journal of the Institute of Convergence Signal Processing
• /
• v.11 no.2
• /
• pp.97-103
• /
• 2010

In this paper, we propose a new digital watermarking technique. The main idea of the proposed algorithm relies on the assumption that any real number can be expressed as a summation of the exponential form with base of 2 and if only consider the first few summations some numbers can be expressed in the same form. Therefore, we can be sure that some amount of changes does not affect the first few summations. The algorithm decomposes a host image in wavelet domain and intensity of the significant wavelet coefficient is expressed in exponential form with base of 2. Multiple barcode watermarks are then embedded by modifying the parity of the exponent. The proposed scheme is semi-blind and also offers either objective or subjective deteew su as well. From extracted watermarks, more accurate watermark is obtained by merging technique as a final watermark. As a simulation result, the proposed algorithm could resist most cases of salt and pepper noise, Gaussian noise and JPEG compression.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.6 no.3
• /
• pp.81-88
• /
• 1996

Cluster analysis is based on partitioning a collection of data points into a number of clusters, where the data points in side a cluster have a certain degree of similarity and it is a fundamental process of data analysis. So, it has been playing an important role in solving many problems in pattern recognition and image processing. For these many clustering algorithms depending on distance criteria have been developed and fuzzy set theory has been introduced to reflect the description of real data, where boundaries might be fuzzy. If fuzzy cluster analysis is tomake a significant contribution to engineering applications, much more attention must be paid to fundamental questions of cluster validity problem which is how well it has identified the structure that is present in the data. Several validity functionals such as partition coefficient, claasification entropy and proportion exponent, have been used for measuring validity mathematically. But the issue of cluster validity involves complex aspects, it is difficult to measure it with one measuring function as the conventional study. In this paper, we propose four performance indices and the way to measure the quality of clustering formed by given learning strategy.