Datasets and Benchmarking

Part of the SpringerBriefs in Computer Science book series (BRIEFSCOMPUTER)


The first pedestrian datasets were aimed at evaluating the new classifiers presented, then they started evaluating the complete detection results. We summarize the existing pedestrian datasets so far, pointing out their most significant aspects. A very important aspect in the evaluation of both classifiers and systems is the protocol used. In this chapter we also review them, describing their interesting points and their use depending on what is evaluated.



Copyright information

© The Author(s) 2014

Authors and Affiliations

  1. 1.Computer Vision CenterUniversitat Autònoma de BarcelonaBellaterraSpain

Personalised recommendations