STEREO- AND NEURAL NETWORK-BASED PEDESTRIAN DETECTION

Pedestrian detection is essential to avoid dangerous traffic situations. The paper presents a fast and robust algorithm for detecting pedestrians in a cluttered urban scene from a pair of moving cameras. This is achieved through stereo-based segmentation and neural network-based recognition. The experiments on a large number of urban street scenes demonstrate that the proposed algorithm: 1) can detect pedestrians in various poses, shapes, sizes, clothing; 2) runs in real time; and 3) is robust to illumination and background changes.

• Supplemental Notes:
  • Full conference proceedings available on CD-ROM.
• Corporate Authors:

  Federal Transit Administration

  1200 New Jersey Avenue, SE
  Washington, DC  United States  20590
• Authors:
  • Zhao, L
  • Thorpe, C E
• Conference:
  • Technical Information on Transit IVI Projects
  • Location: Las Vegas, Nevada
  • Date: 2002-10-28 to 2002-10-29
• Publication Date: 2002

Language

• English

Media Info

• Pagination: 7p

Subject/Index Terms

• TRT Terms: Cameras; Image processing; Intelligent vehicles; Neural networks; Pedestrian detectors; Pedestrian safety; Safety; Stereoscopic cameras
• Identifier Terms: Intelligent Vehicle Initiative
• Subject Areas: Operations and Traffic Management; Public Transportation; Safety and Human Factors;

Filing Info

• Accession Number: 00936274
• Record Type: Publication
• Files: TRIS, USDOT
• Created Date: Jan 3 2003 12:00AM