Sorted by DateClassified by Publication TypeSorted by First Author Last NameClassified by Topic

Fast Human Detection for Indoor Mobile Robots Using Depth Images

Benjamin Choi, Çetin Meriçli, Joydeep Biswas, and Manuela Veloso. Fast Human Detection for Indoor Mobile Robots Using Depth Images. In IEEE International Conference on Robotics and Automation (ICRA), 2013. (accepted)

Download

[PDF] 

Abstract

A human detection algorithm running on an indoor mobile robot has to addresschallenges including occlusions due to cluttered environments, changingbackgrounds due to the robot's motion, and limited on-board computationalresources. We introduce a fast human detection algorithm for mobile robots equipped withdepth cameras. First, we segment the raw depth image using a graph-basedsegmentation algorithm. Next, we apply a set of parameterized heuristics tofilter and merge the segmented regions to obtain a set of candidates. Finally,we compute a Histogram of Oriented Depth (HOD) descriptor for eachcandidate, and test for human presence with a linear SVM. We experimentallyevaluate our approach on a publicly available dataset of humans in an open areaas well as our own dataset of humans in a cluttered cafe environment. Ouralgorithm performs comparably well on a single CPU core against anotherHOD-based algorithm that runs on a GPU even when the number of training examplesis decreased by half. We discuss the impact of the number of training examples on performance, anddemonstrate that our approach is able to detect humans in different postures (e.g., standing, walking, sitting) and with occlusions.

BibTeX

@inproceedings{choi2013a,
  author    = {Benjamin Choi and Çetin Meriçli and Joydeep Biswas and Manuela Veloso},
  title     = {Fast Human Detection for Indoor Mobile Robots Using Depth Images},
  booktitle = {IEEE International Conference on Robotics and Automation (ICRA)},
  year      = {2013},
  abstract  = {A human detection algorithm running on an indoor mobile robot has to address
challenges including occlusions due to cluttered environments, changing
backgrounds due to the robot's motion, and limited on-board computational
resources. We introduce a fast human detection algorithm for mobile robots equipped with
depth cameras. First, we segment the raw depth image using a graph-based
segmentation algorithm. Next, we apply a set of parameterized heuristics to
filter and merge the segmented regions to obtain a set of candidates. Finally,
we compute a Histogram of Oriented Depth (HOD) descriptor for each
candidate, and test for human presence with a linear SVM. We experimentally
evaluate our approach on a publicly available dataset of humans in an open area
as well as our own dataset of humans in a cluttered cafe environment. Our
algorithm performs comparably well on a single CPU core against another
HOD-based algorithm that runs on a GPU even when the number of training examples
is decreased by half. We discuss the impact of the number of training examples on performance, and
demonstrate that our approach is able to detect humans in different postures (e.g., standing, walking, sitting) and with occlusions.},
  note      = {(accepted)},
  bib2html_pubtype = {Refereed Conference},
  bib2html_rescat = {Computer Vision, Human-Robot Interaction},
  bib2html_dl_pdf = {../files/choiICRA2013HumanDetection.pdf},
}

Generated by bib2html.pl (written by Patrick Riley ) on Thu May 01, 2014 16:27:43