Efficient Pedestrian Detection Using Multi-Scale HOG Features with Low Computational Complexity

Soojin KIM  Kyeongsoon CHO  

Publication
IEICE TRANSACTIONS on Information and Systems   Vol.E97-D   No.2   pp.366-369
Publication Date: 2014/02/01
Online ISSN: 1745-1361
DOI: 10.1587/transinf.E97.D.366
Print ISSN: 0916-8532
Type of Manuscript: LETTER
Category: Pattern Recognition
Keyword: 
pedestrian detection,  interpolation,  multi-scale HOG,  integral HOG,  two-stage approach,  

Full Text: PDF(577.2KB)
>>Buy this Article


Summary: 
In this paper, an efficient method to reduce computational complexity for pedestrian detection is presented. Since trilinear interpolation is not used, the amount of required operations for histogram of oriented gradient (HOG) feature calculation is significantly reduced. By calculating multi-scale HOG features with integral HOG in a two-stage approach, both high detection rate and speed are achieved in the proposed method.