Title :
Multi-scale bidirectional local template patterns for real-time human detection
Author :
Jiu Xu ; Ning Jiang ; Xinwei Xue ; Heming Sun ; Wenxin Yu ; Goto, Satoshi
Author_Institution :
Grad. Sch. of IPS, Waseda Univ., Kitakyushu, Japan
fDate :
Sept. 30 2013-Oct. 2 2013
Abstract :
In this paper, a feature named multi-scale bidirectional local template patterns (MBLTP) is proposed for human detection. As an extension of bidirectional local template patterns (BLTP), MBLTP not only integrates the textural and gradient information according to the four predefined templates but also calculates information for additional feature vectors by adjusting the scale of the training samples. These additional feature vectors contain multi-scale information on the samples, which can make the feature more discriminative than its original form. Experimental results for an INRIA dataset show that the detection rate of our proposed MBLTP feature outperforms those of other features such as the multi-level histogram of orientated gradient (multi-level HOG), multi scale block histogram of template (MB-HOT), and HOG-LBP. Moreover, in order to make our feature meet real-time requirements, an implementation based on a graphic process unit (GPU) is adopted to accelerate the calculation.
Keywords :
feature extraction; gradient methods; graphics processing units; image texture; object detection; GPU; HOG-LBP; INRIA dataset; MB-HOT; MBLTP; feature vectors; gradient information; graphic process unit; multilevel HOG; multilevel histogram of orientated gradient; multiscale bidirectional local template patterns; multiscale block histogram of template; multiscale information; real-time human detection; textural information; Graphics processing units; Histograms; Sun; Vectors;
Conference_Titel :
Multimedia Signal Processing (MMSP), 2013 IEEE 15th International Workshop on
Conference_Location :
Pula
DOI :
10.1109/MMSP.2013.6659318