Exploring AI-Based Video Segmentation and Saliency Computation To Optimize Imagery-Acquisition From Moving Vehicles

Details

• Creators:
  Silva, Claudio T ; Ozbay, Kaan ; Rulff de Costa, Joao ; Lin, Jianzhe ; Hosseini, Maryam ; Tokuda, Eric
• Corporate Creators:
  Connected Cities for Smart Mobility toward Accessible and Resilient Transportation Center (C2SMART) ; New York University. Tandon School of Engineering
• Corporate Contributors:
  United States. Department of Transportation. University Transportation Centers (UTC) Program ; United States. Department of Transportation. Office of the Assistant Secretary for Research and Technology ; United States. Department of Transportation. Federal Highway Administration
• Subject/TRT Terms:
  • Artificial intelligence
  • Computer vision
  • Data collection
  • Image processing
  • Machine learning
  • Pedestrian counts
  • Sensors
  • Video
  • Infrastructure
  • StreetAware
• Resource Type:
  Tech Report
• Geographical Coverage:
  United States
• Edition:
  Final report, 3/1/21-10/31/23
• Corporate Publisher:
  Connected Cities for Smart Mobility toward Accessible and Resilient Transportation Center (C2SMART)
• Abstract:
  In this study, a new dataset and tool were generated describing and mapping street-level infrastructure. The presented dataset, StreetAware, is generated from more than 7 hours of synchronized data collected at urban intersections by specialized Reconfigurable Environmental Intelligence Platform (REIP) sensors developed by the Visualization and Data Analytics (VIDA) Research Center at NYU. To demonstrate these key features of the data, we present four uses for the data that are not possible on many existing datasets. (1) to track objects using the multiple perspectives of multiple cameras from both audio (sound-based localization) and visual modes, (2) to associate audio events with their respective visual representations using audio and video, (3) to track the amount of each type of object in a scene over time, i.e., occupancy, and (4) to measure the speed of a pedestrian while crossing a street using multiple synchronized views and the high-resolution capability of the cameras.
• Format:
  PDF
• Funding:
  69A3551747119
• Download URL:
  https://rosap.ntl.bts.gov/view/dot/72505/dot_72505_DS1.pdf
• File Type:
  [PDF - 2.32 MB]
• Collection(s):
  Artificial Intelligence Deployment Center of Excellence (AI COE) ; University Transportation Centers
• Main Document Checksum:
  urn:sha-512:57885e4e9270f4a0850b9543fd1541b43fa78f8412ccd81b79d1ca6ac26da41e01a33fecf26522991db0bc0eb59e03233459de6f39e353c2441762db39c0ae61