2014, Vol.19, No.6

Journal of The Korea Society of Computer and Information 2023 KCI Impact Factor : 0.65

pISSN : 1598-849X / eISSN : 2383-9945
https://journal.kci.go.kr/jksci

In this paper, we propose a real-time video playback scheme for the N-Screen service based onWindows Azure. This scheme creates several playback blocks based on the performance of eachnode by non-uniform splitting of the original video. To reduce transcoding-time, it allocates the playback blocks to a corresponding node by transcoding the playback blocks. Through thesimulation, we show that it is more effective to use real-time video playback for the N-screenservice than the previous method. The proposed scheme splits an AVI format 300MB source videowith non-uniform playback blocks. It allocates the playback blocks to the heterogeneous node ofWindows Azure, the commercial cloud system and measures of transcoding-time by transcodingnon-uniform playback blocks to mp4 and Flv format. As a result, the proposed scheme improvesthe performance of the N-screen service based on Windows Azure compared to the previousuniform split strategy.

While safe and convenient, ultrasound imaging analysis is often criticized by its subjectivedecision making nature by field experts in analyzing musculoskeletal system. In this paper, wepropose a new automatic method to extract muscle area using ART2 neural network basedquantization. A series of image processing algorithms such as histogram smoothing and End-insearch stretching are applied in pre-processing phase to remove noises effectively. Muscle areasare extracted by considering various morphological features and corresponding analysis. Inexperiment, our ART2 based Quantization is verified as more effective than other generalquantization methods.

An ensemble classifier system is a widely-used multi-classifier system, which combines theresults from each classifier and, as a result, achieves better classification result than any singleclassifier used. Several methods have been used to build an ensemble classifier including boosting,which is a cascade method where misclassified examples in previous stage are used to boost theperformance in current stage. Boosting is, however, a serial method which does not form acomplete feedback loop. In this paper, proposed is context sensitive SVM ensemble (CASE) whichadopts SVM, one of the best classifiers in term of classification rate, as a basic classifier andclustering method to divide feature space into contexts. As CASE divides feature space and trains SVMs simultaneously, the result from one component can be applied to the other and CASEachieves better result than boosting. Experimental results prove the usefulness of the proposedmethod.