A study on a method for stable pedestrian detection against pose changes with generative learning

Recently, pedestrian detection from in-vehicle camera images is being focused. However, it is difficult to detect pedestrians due to the variety of their poses and backgrounds. To tackle this problem, we propose a method to detect various pedestrians from in-vehicle camera images. To deal with changes of pedestrians’ pose and environment, most existing methods making use of their appearance require to prepare a lot of pedestrian images manually. The proposed method classifies a small number of pedestrian images into several pose classes and then generates various pedestrian images from each pose class. Finally, the proposed method constructs a classifier based on multiple templates from each pedestrian pose. Experimental results showed that the detection accuracy of the method outperformed existing methods, and we confirmed its effectiveness. [Note]This document is an informal handout distributed at an IEICE TC-PRMU workshop.

[1]  Timothy F. Cootes,et al.  Active Shape Models-Their Training and Application , 1995, Comput. Vis. Image Underst..

[2]  Hiroshi Murase,et al.  Recognition of Road Markings from In-Vehicle Camera Images by a Generative Learning Method , 2009, MVA.

[3]  Bill Triggs,et al.  Histograms of oriented gradients for human detection , 2005, 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05).

[4]  Chih-Jen Lin,et al.  LIBSVM: A library for support vector machines , 2011, TIST.

[5]  Dariu Gavrila,et al.  Virtual sample generation for template-based shape matching , 2001, Proceedings of the 2001 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. CVPR 2001.

[6]  Dariu Gavrila,et al.  A mixed generative-discriminative framework for pedestrian classification , 2008, 2008 IEEE Conference on Computer Vision and Pattern Recognition.

[7]  David Vázquez,et al.  Learning appearance in virtual scenarios for pedestrian detection , 2010, 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition.