Parallelization of Face Detection using OpenMP

Aditya Mitra, Anuj Chaudhary, Riddhi Maniktalia, Kushaj Arora

Abstract: In OpenCV, all predefined libraries and functions are C - codes that run in serial fashion. Similar frameworks are also available for CUDA, however using CUDA is only possible if you have an NVIDIA graphics card, which is obviously not the case for everyone. OPENMP?s release of a stable version in late 2015 facilitated parallel processing and broadened its developer base. Because OPENMP could run in parallel on typical computers with numerous cores and was portable, it quickly gained popularity. To parallelize a face identification algorithm, we will use C++ and OpenMP. To better understand the findings, we will compare the results with a serial code. The time it took the machine to process the image in serial code compared to the time it took the identical computer to process the image in parallel code would be the factors for determining the outcomes.

Keywords: OpenMP, Face Detection, Parallelization, HAAR Filter, Viola Jones Algorithm