سيدعلي اصغر سيدي

Boundary detection in binary images is an important operation in many computer vision applications. The problem is finding boundaries in a 2D binary image (where all pixels are either 0 or 1). Several algorithms are available for this problem, but most of them are ordinal. A parallel boundary tracking algorithm has been proposed, which was designed to run in graphics processing units and was based on a sequential algorithm known as the Suzuki algorithm. In this thesis, we adapt the GPU algorithm so that it can be run on multi-core computers. The resulting algorithm is eva‎luated against its GPU counterpart. The results show that the performance of the GPU algorithm deteriorates (or even fails) for very large images or images with large margins. On the other hand, the proposed multi-core algorithm can effectively cope with large images.

رديابي مرز ، بينايي كامپيوتري ، محاسبات موازي ، GPU ، OpenMP ، محاسبات چند هسته اي

Border tracking · Computer vision · Parallel computing · GPU computing · OpenMP · Multicore computing

seyed aliasghasr seyedi

dr javad vahidi