Diagnosing error in object detectors

Derek Hoiem, Yodsawalai Chodpathumwan, Qieyun Dai

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

This paper shows how to analyze the influences of object characteristics on detection performance and the frequency and impact of different types of false positives. In particular, we examine effects of occlusion, size, aspect ratio, visibility of parts, viewpoint, localization error, and confusion with semantically similar objects, other labeled objects, and background. We analyze two classes of detectors: the Vedaldi et al. multiple kernel learning detector and different versions of the Felzenszwalb et al. detector. Our study shows that sensitivity to size, localization error, and confusion with similar objects are the most impactful forms of error. Our analysis also reveals that many different kinds of improvement are necessary to achieve large gains, making more detailed analysis essential for the progress of recognition research. By making our software and annotations available, we make it effortless for future researchers to perform similar analysis.

Original languageEnglish (US)
Title of host publicationComputer Vision, ECCV 2012 - 12th European Conference on Computer Vision, Proceedings
Pages340-353
Number of pages14
EditionPART 3
DOIs
StatePublished - 2012
Event12th European Conference on Computer Vision, ECCV 2012 - Florence, Italy
Duration: Oct 7 2012Oct 13 2012

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
NumberPART 3
Volume7574 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other12th European Conference on Computer Vision, ECCV 2012
Country/TerritoryItaly
CityFlorence
Period10/7/1210/13/12

ASJC Scopus subject areas

  • Theoretical Computer Science
  • General Computer Science

Fingerprint

Dive into the research topics of 'Diagnosing error in object detectors'. Together they form a unique fingerprint.

Cite this