Morphometric studies of genus Placocheilus （Teleostei： Cypriniformes） from Red River, China

It is practically difficult to differentiate Placocheilus robustus and Placocheilus caudofasciatus from Red River drainage of China. Without stated reasons, P. robustus has been assumed as the synonyms of P. caudofasciatus. The present study aims to decipher the morphological differences between two species so as to provide reliable clues for their classification by multivariate morphometry. A total of 72 specimens of two species in genus Placocheilus were examined. Besides morphological character comparisons, 10 anatomic landmarks were utilized and 23 frame structures and 15 general characters measured. The scatter plot results of principal component analysis showed that all specimens were clustered together and could not be defined as two distinct species. No significant morphological differences existed in four diagnostic characters between P. robustus and P. caudofasciatus. Thus the results of the present study fail to support P. robustus as a valid and independent species.

Application of fuzzy clustering method to determining sub-fault planes of earthquake from aftershocks sequence

Earthquake rupture process generally involves several faults activities instead of a single fault A new method using both fuzzy clustering and principal component analysis makes it possible to reconstruct three dimensional structure of involved faults in earthquake if the aftershocks around the active fault planes distribute uniformly. When seismic events are given, the optimal faults structures can be determined by our new method. Each of sub-fault planes is fully characterized by its central location, length, width, strike and dip. The resolution determines the number of fault segments needed to describe the earthquake catalog. The higher the resolution, the finer the structure of the reconstructed fault segments. The new method successfully reconstructs the fault segments using synthetic earthquake catalogs. By taking the 28 June 1992 Landers earthquake oceured in southern California as an example, the reconstructed fault segments are consistent with the faults already known on geological maps or blind faults that appeared quite frequently in longer-term catalogs.

OPTIMIZATION STUDY IN MANUFACTURING PROCESS FOR PC/ABS BLENDS

The optimization of injection molding process for polycarbonate/acrylonitrile-butadiene-styrene (PC/ABS) blends is studied using Taguchi method and principalcomponent analysis (PCA). Four controllable process factors are studied at three levels each in themanufacturing process. The L_9 orthogonal array is conducted to determine the optimum processfactor/level combination for single quality of mechanical properties. In addition, the principalcomponent analysis is employed to trans-form the correlated mechanical properties to a set ofuncorrelated components and to evaluate a comprehensive index for multi-response cases. Then theoptimum process factor/level combination for multiple qualities can be determined. Finally, theanalysis of variance is used to find out the most influential injection molding parameter for singleand multiple qualities problems.

Parametric shape prior model used in image segmentation

Due to the frequency of occlusion, cluttering and lowcontrast edges, gray intensity based active contour models oftenfail to segment meaningful objects. Prior shape information is usuallyutilized to segment desirable objects. A parametric shape priormodel is proposed. Firstly, principal component analysis is employedto train object shape and transformation is added to shaperepresentation. Then the energy function is constructed througha combination of shape prior energy, gray intensity energy andshape constraint energy of the kernel density function. The objectboundary extraction process is converted into the parameters solvingprocess of object shape. Besides, two new shape prior energyfunctions are defined when desirable objects are occluded by otherobjects or some parts of them are missing. Finally, an alternatingdecent iteration solving scheme is proposed for numerical implementation.Experiments on synthetic and real images demonstratethe robustness and accuracy of the proposed method.

Spatial variations of river water quality in Pearl River Delta, China

In Pearl River Delta （PRD）, river water quality has deteriorated gradually due to population increase and ongoing industrialization and urbanization. In this study, multivariate statistic methods were used to assess water quality spatial pattern and to identify characteristics of water quality variation in the PRD. Water quality monitoring of the PRD during the year 2005 and 2008 was conducted at 25 different stations. Seventeen water quality parameters were analyzed for further studying. Results of one-way analysis of variance （ANOVA） indicated that all the parameters except air temperature, water temperature and zinc showed significant difference among monitoring stations in both dry and wet season. Monitoring stations in the PRD were separately classified into three statistically significant clusters at （Olink/Omax） （ 2 in dry and wet season, respectively. The three clusters indicated the similarity and dissimilarity of river water quality among 25 monitoring stations, corresponding to heavy pollution, moderate pollution and slight pollution. Thus, the results of this study are useful to evaluate water quality and manage water resources in the PRD.

A statistical model to predict total column ozone in Peninsular Malaysia

This study aims to predict monthly columnar ozone in Peninsular Malaysia based on concentrations of several atmospheric gases. Data pertaining to five atmo- spheric gases （CO2, 03, CH4, NO2, and H2O vapor） were retrieved by satellite scanning imaging absorption spectro- metry for atmospheric chartography from 2003 to 2008 and used to develop a model to predict columnar ozone in Peninsular Malaysia. Analyses of the northeast monsoon （NEM） and the southwest monsoon （SWM） seasons were conducted separately. Based on the Pearson correlation matrices, columnar ozone was negatively correlated with HzO vapor but positively correlated with COz and NO2 during both the NEM and SWM seasons from 2003 to 2008. This result was expected because NO/is a precursor of ozone. Therefore, an increase in columnar ozone concentration is associated with an increase in NO2 but a decrease in H/O vapor. In the NEM season, columnar ozone was negatively correlated with H20 （-0.847）, NO2 （0.754）, and CO2 （0.477）; columnar ozone was also negatively but weakly correlated with CH4 （-0.035）. In the SWM season, columnar ozone was highly positively correlated with NO2 （0.855）, CO2 （0.572）, and CH4（0.321） and also highly negatively correlated with H2O（-0.832）. Both multiple regression and principal component analyses were used to predict the columnar ozone value in Peninsular Malaysia. We obtained the best-fitting regression equations for the columnar ozone data using four independent variables. Our results show approxi- mately the same R value （≈0.83） for both the NEM and SWM seasons.

A pilot study for distinguishing basal cell carcinoma from normal human skin tissues using visible resonance Raman spectroscopy

Aim: The aim of the study is to test visible resonance Raman (VRR) spectroscopy for rapid skin cancer diagnosis,and evaluate its effectiveness as a new optical biopsy method to distinguish basal cell carcinoma (BCC) from normal skin tissues.Methods: The VRR spectroscopic technique was undertaken using 532 nm excitation. Normal and BCC human skin tissue samples were measured in seconds. The molecular fingerprints of various native biomolecules as biomarkers were analyzed. A principal component analysis - support vector machine (PCA-SVM) statistical analysis method based on the molecular fingerprints was developed for differentiating BCC from normal skin tissues.Results: VRR provides a rapid method and enhanced Raman signals from biomolecules with resonant and nearresonant absorption bands as compared with using a near-infrared excitation light source. The VRR technique revealed chemical composition changes of native biomarkers such as tryptophan, carotenoids, lipids and proteins.The VRR spectra from BCC samples showed a strong enhancement in proteins including collagen type I combined with amide I and amino acids, and a decrease in carotenoids and lipids. The PCA-SVM statistical analysis based on the molecular fingerprints of the biomarkers yielded a 93.0% diagnostic sensitivity, 100% specificity, and 94.5%accuracy compared with histopathology reports.Conclusion: VRR can enhance molecular vibrational modes of various native biomarkers to allow for very fast display of Raman modes in seconds. It may be used as a label-free molecular pathology method for diagnosis of skin cancer and other diseases and be used for combined treatment with Mohs surgery for BCC.