Po-Hsun Tseng

* Developed a scalable parallel algorithm for polygon operations using Quadtree and graph partitioning. Successfully scaled to 32 compute nodes with a 0.2 load-balance index and 71% parallel efficiency. * Implemented the Hopkins’ method for computational lithography using OpenMP.

* Designed a sweep line algorithm combined with a self-balancing interval tree and the Red-Black tree forest, achieving optimal O(N(k+LogN)) time complexity for counting overlapping polygons on a design file. N total number of rectangles, k reported polygons. * Implemented Bilateral filter for noise reduction in CUDA, utilizing shared memory for efficiency. * Enhanced image stacking throughput using CUDA, achieving noise reduction and increasing speed by a factor of N. N is the number of image pixels.