ABSTRACT
Embedding (TSPE) .......................................................... 294 8.2.3 Quality Of Visualization ................................................... 295 8.3 Remote Sensing Imagery ............................................................. 297 8.4 Our Contributions in GPU Application........................................ 298 8.4.1 Parallel TSPE ................................................................... 298
8.4.2 Parallel Correlation Metric ............................................... 299 8.4.3 Parallel Residual Variance Metric .................................... 300 8.5 Experimental Results ................................................................... 302 8.6 Conclusion ................................................................................... 304 Keywords .............................................................................................. 305 References ............................................................................................. 305
In this chapter, we present a parallel method to visualize remote sensing imagery data sets and measure their efficiency on the graphics processing unit (GPU). Visualization of remote sensing imagery data sets is a common challenge task in the dimensionality reduction (DR). The requirement to accelerate the projection process and efficiency measurement of the visualization comes from the large size of the data sets. We have implemented the trustworthy stochastic proximity embedding (TSPE) method on GPU to speed up its projection process. To measure the efficiency of the visualization, the parallel codes of the two wellknown metrics in this field namely, correlation and residual variance are introduced. The results showed that the high computational efficiency of the GPU helped to reduce the time spent on processing the results and computing their efficiency.
