Geodesic
Integer Arithmetic Approximation of the HoG Algorithm used for Pedestrian Detection
Computer Science and Information Systems, Tome 14 (2017) no. 2.

Voir la notice de l'article provenant de la source Computer Science and Information Systems website

This paper presents the results of a study of the effects of integer (fixed-point) arithmetic implementation on classification accuracy of a popular open-source people detection system based on Histogram of Oriented Gradients. It is investigated how the system performance deviates from the reference algorithm performance as integer arithmetic is introduced with different bit-width in several critical parts of the system. In performed experiments, the effects of different bit-width integer arithmetic implementation for four key operations were separately considered: HoG descriptor magnitude calculation, HoG descriptor angle calculation, normalization and SVM classification. It is found that a 13-bit representation of variables is more than sufficient to accurately implement this system in integer arithmetic. The experiments in the paper are conducted for pedestrian detection and the methodology and the lessons learned from this study allow generalization of conclusions to a broader class of applications.
Keywords: computer vision, fixed-point, histogram of oriented gradients, pedestrian detection 
@article{CSIS_2017_14_2_a3,
     author = {Sr{\dj}an Sladojevi\'c and Andra\v{s} Anderla and Dubravko \'Culibrk and Darko Stefanovi\'c1 and Bojan Lali\'c},
     title = {Integer {Arithmetic} {Approximation} of the {HoG} {Algorithm} used for {Pedestrian} {Detection}},
     journal = {Computer Science and Information Systems},
     publisher = {mathdoc},
     volume = {14},
     number = {2},
     year = {2017},
     url = {http://geodesic.mathdoc.fr/item/CSIS_2017_14_2_a3/}
}
TY  - JOUR
AU  - Srđan Sladojević
AU  - Andraš Anderla
AU  - Dubravko Ćulibrk
AU  - Darko Stefanović1
AU  - Bojan Lalić
TI  - Integer Arithmetic Approximation of the HoG Algorithm used for Pedestrian Detection
JO  - Computer Science and Information Systems
PY  - 2017
VL  - 14
IS  - 2
PB  - mathdoc
UR  - http://geodesic.mathdoc.fr/item/CSIS_2017_14_2_a3/
ID  - CSIS_2017_14_2_a3
ER  - 
%0 Journal Article
%A Srđan Sladojević
%A Andraš Anderla
%A Dubravko Ćulibrk
%A Darko Stefanović1
%A Bojan Lalić
%T Integer Arithmetic Approximation of the HoG Algorithm used for Pedestrian Detection
%J Computer Science and Information Systems
%D 2017
%V 14
%N 2
%I mathdoc
%U http://geodesic.mathdoc.fr/item/CSIS_2017_14_2_a3/
%F CSIS_2017_14_2_a3
Srđan Sladojević; Andraš Anderla; Dubravko Ćulibrk; Darko Stefanović1; Bojan Lalić. Integer Arithmetic Approximation of the HoG Algorithm used for Pedestrian Detection. Computer Science and Information Systems, Tome 14 (2017) no. 2. http://geodesic.mathdoc.fr/item/CSIS_2017_14_2_a3/