Rights accessRestricted access - confidentiality agreement
This master-thesis proposes a novel method for long range pedestrian
detection in outdoor scenarios from static images, where pedestrians can
appear in crowded scenes and with different kinds of occlusion. An
efficient way to extract high gradient information edges from an image
using its thresholded Gradient Magnitude Image and the segmentation of
the image in question is presented.
Furthermore, three new kinds of features have been designed, which are
extracted from the high information edges: HIST, VEC and HIST VEC
features. The results show that the third type (HIST VEC features), which is
a combination of the two first (HIST and VEC features), gives the best
Although there is not any published work focused only on detection of
long distance pedestrians from static images to compare, the presented
results point out the usefulness of this approach.
All rights reserved. This work is protected by the corresponding intellectual and industrial property rights. Without prejudice to any existing legal exemptions, reproduction, distribution, public communication or transformation of this work are prohibited without permission of the copyright holder. If you wish to make any use of the work not provided for in the law, please contact: email@example.com