RESUMO
A novel method was developed to classify hyperspectral remote sensing image based on independent component analysis (ICA) and support vector machine (SVM) algorithms. The characteristic information of the hyperspectral remote sensing image captured by PHI (made in China, with 80 bands) was extracted by ICA algorithm, and SVM classifier was established with the extracted image data (20 spectral dimensions). After kernel function selecting and parameter optimizing, it was found that the SVM algorithm(RBF kernel function; parameter C = 1093), gamma = 0.05) with accuracy 94.5127% and kappa coefficient 0.9351 has the best classification result, better than the results of four kinds of conventional algorithms, including neural net classification (accuracy 39.4758% and kappa coefficient 0.3155), spectral angle mapper classification (accuracy 80.2826% and kappa coefficient 0.7709), minimum distance classification (accuracy 85.4627% and kappa coefficient 0.8277) and maximum likelihood classification (accuracy 86.0156% and Kappa coefficient 0.8351). In order to control the "pepper and salt" phenomenon which appeared in classification map frequently, the classification result of SVM (RBF kernel) was operated by the method of clump classes using the morphological operators, and that the classification map closer to actual situation was acquired, with the accuracy and kappa coefficient increasing to 94.7584% and 0.9380, respectively. The study indicated that the ICA combined with SVM was an preferred method for hyperspectral remote sensing image classification, and clump classes was a effective method to optimized the classification result.
RESUMO
Epstein-Barr virus (EBV) infection is associated with salivary gland lymphoepithelial carcinoma (SLEC) and nasopharyngeal carcinoma (NPC). EBV is a ubiquitous herpes virus world wide, but EBV-associated SLEC and NPC are prevalent in restricted regions such as south areas of China, Southeastern Asia and Greenland (Eskimos). To examine whether particular EBV variants play roles in the development of SLEC and NPC, we isolated the complete EBV LMP1 genes from 12 paraffin-embedded biopsy samples of SLECs isolated from China, Taiwan and Russia, and compared these LMP1 genes with those of NPC (CAO) and the prototype B95-8 EBV. Nucleotide sequence analysis showed that SLECs LMP1 is more similar to that of CAO than that of prototype B95-8. The analysis also identified several conserved (67-100%) variations in SLEC-LMP1 and CAO-LMP1 distinct from B95-8-LMP1. These included 10-amino acid deletion, 5-amino acid deletion and 12-single amino acid variations. A SLEC-LMP1 gene with the aforementioned conserved variations inhibited the growth of an embryonic kidney cell line (293T), highly activated the NF-kappaB pathway, and these activities were equivalent to those of B95-8 and CAO. These findings suggest that the biological functions of SLEC-LMP 1 are similar to those of B95-8-LMP1 and CAO-LMP1, and that these amino acid variations including the well-known 10-aa deletion did not affect these two prominent activities. While the present results could not uncover functional differences between SLEC-LMP1 and B95-8-LMP1, the nucleotide sequences and the molecular clone of LMP1 directly isolated from SLEC patients will be a useful tool to identify the high-pathogenic EBV strain(s), associated with SLEC and NPC.