Title :
Face detection algorithm using haar-like feature for GPU architecture
Author :
Pertsau, Dmitry ; Uvarov, A.
Author_Institution :
Belarusian State Univ. of Inf. & Radioelectron., Minsk, Belarus
Abstract :
This article describes parallel algorithm of face detection on images for GPU architecture. This algorithm is an extension of an algorithm from OpenCV library. A computational structure is presented for the developed algorithm. Also, scheduling algorithm was developed to balance a workload among GPU´s threads.
Keywords :
Haar transforms; face recognition; graphics processing units; object detection; parallel algorithms; resource allocation; scheduling; GPU architecture; GPU threads; Haar-like feature; OpenCV library; computational structure; face detection algorithm; parallel algorithm; scheduling algorithm; workload balancing; Classification algorithms; Face detection; Feature extraction; Graphics processing units; Instruction sets; Kernel; Real-time systems; CUDA; GPU; Haar-like features; OpenCV;
Conference_Titel :
Intelligent Data Acquisition and Advanced Computing Systems (IDAACS), 2013 IEEE 7th International Conference on
Conference_Location :
Berlin
Print_ISBN :
978-1-4799-1426-5
DOI :
10.1109/IDAACS.2013.6663020
