The DTW Data Distribution over a Grid Computing Architecture

In this paper; we show how grid computing can further speedup Dynamic Time Warping (DTW) algorithm. More specifically, we present experimental results of the data distribution of the Arabic printed cursive OCR based on the Dynamic Time Warping (DTW) algorithm over the Scientific Research Tunisian Grid (SRTG). In fact, the Arabic printed cursive OCR based on the DTW algorithm provides very interesting recognition and segmentation rates. Among the advantages of the DTW algorithm, is its ability to achieve properly and simultaneously the recognition and the segmentation of connected or cursive characters from within a reference library of isolated characters. Unfortunately, the big amount of computing to be achieved during the recognition process makes the DTW execution time very slow and hence restricts its utilization. Conducted experiments show and confirm that grid computing presents a very interesting framework to speedup the DTW execution time. In addition, we found that the speedup and the efficiency factors increase with the size of the Arabic text.