Abstract:
Nowadays, with the introduction of Graphics Processing Unit for general purpose
issues, not only graphical ones, there has been an increasing attention towards exploiting
GPU processing power for deep learning algorithms. In the world of technology as
considered by many scientist and analysts everything is going very fast.In this thesis I
will use FPGA boards rather than graphical card processing using like NVidia, as a case
study in observation of behavior regarding with image segmentation.
Therea are several qualities that distinguish both processors, with classical graphical
processing units being more flexible, not much complex and vice versa with FPGA
processors offering programmability, reducing time latency, optimized energy in
computation process. For while, it has been an enigma the comparison between two
different mentalities ( software vs. hardware ) engineering mentalities will occur, thus
the results will be compared to each-other like energy used, flip-flops, LUT-s etc. The
whole system will be implemented in Xilinx Nexus A7 FPGA board and Vivado HLS
2018.2 software framework.
