Human Face Detection in Digital Video Using SVMEnsemble

This Letter proposes automatic human face detection in digital video using a support vector machine (SVM) ensemble to improve the detection performance. The SVM ensemble consists of several independently trained SVMs using randomly chosen training samples via a bootstrap technique. Next, they are aggregated in order to make a collective decision via a majority voting scheme. Experimental results show that the proposed face detection method using SVM ensemble outperforms conventional methods such as using only single SVM and Multi-Layer Perceptron in terms of classification accuracy, false alarms, and missing rates.     

Constructing support vector machine ensemble

Even the support vector machine (SVM) has been proposed to provide a good generalization performance, the classification result of the practically implemented SVM is often far from the theoretically expected level because their implementations are based on the approximated algorithms due to the high complexity of time and space. To improve the limited classification performance of the real SVM, we propose to use the SVM ensemble with bagging (bootstrap aggregating) or boosting. In bagging, each individual SVM is trained independently using the randomly chosen training samples via a bootstrap technique. In boosting, each individual SVM is trained using the training samples chosen according to the sample's probability distribution that is updated in proportional to the errorness of the sample. In both bagging and boosting, the trained individual SVMs are aggregated to make a collective decision in several ways such as the majority voting, least-squares estimation-based weighting, and the double-layer hierarchical combining. Various simulation results for the IRIS data classification and the hand-written digit recognition, and the fraud detection show that the proposed SVM ensemble with bagging or boosting outperforms a single SVM in terms of classification accuracy greatly.     

National radon survey in Korea

To estimate annual average concentrations in Korean dwellings and the effective dose to the general public, nationwide surveys on radon were conducted in 1989, 1999-2000 and 2002-2005. The total number of dwellings was about 5600. A survey of thoron and its decay products was also conducted in 2002-2005. In 2008-2009, a new radon survey in 1100 public buildings was conducted. The annual arithmetic (AM) and geometric (GM) means of indoor radon concentration in total were 62.1 ± 66.4 and 49.0 ± 1.9 Bq m(-3), respectively. The annual AM and GM means of indoor thoron concentrations were 40.4 ± 56.0 and 10.7 ± 2.9 Bq m(-3), respectively. The radon and thoron concentrations in detached houses were much higher than those in apartments. The locations of the high radon or thoron houses seem to be correlated with the concentrations of their parent nuclides in surface soil. The mean individual doses of radon and thoron were calculated to be 1.65 and 0.17 mSv y(-1), respectively.     

Synthesis of CdS nanoparticles dispersed within solutions and polymer films using amphiphilic urethane acrylate chains

In this study, a new process for preparation of CdS/polymer nanocomposite films using a new amphiphilic oligomer chain (Urethane Acrylate Nonionomer: UAN) is presented. UAN chains have hydrophilic polyethylene oxide segments and polypropylene oxide-based hydrophobic segments as part of the same backbone. In addition, these chains also have reactive vinyl groups at their hydrophobic segments. For the preparation of CdS nano-colloid solutions, UAN chains acted as a stabilizer for CdS nanoparticles dispersed in a solvent. For the fabrication of CdS/polymer nanocomposite films, UAN chains in CdS nano-colloid solutions were polymerized through reactions between their vinyl groups, resulting in the formation of CdS nanoparticle dispersed films swollen by the solvent. The nature of the solvent used strongly influenced the size of the CdS nanoparticles, which was confirmed by UV absorption spectra, PL emission spectra and TEM images. Smaller sized CdS nanoparticles were formed in higher polar solvents such as DMAc and methanol, which can be explained by the higher solubility of UAN chains and the more effective dissociation and stabilization of cadmium salts and CdS nanoparticles by UAN chains in the polar solvent.