Advanced Search

Browse by Discipline

Scientific Societies

E-print Alerts

Add E-prints

E-print Network

  Advanced Search  

Performance Measures for Object Detection Evaluation Bahadir Ozdemira

Summary: Performance Measures for Object Detection Evaluation
Bahadir ¨Ozdemira
, Selim Aksoy,a
, Sandra Eckertb
, Martino Pesaresib
, Daniele Ehrlichb
aDepartment of Computer Engineering, Bilkent University, Ankara, 06800, Turkey
bInstitute for the Protection and Security of the Citizen, European Commission, Joint Research Centre, 21020 Ispra (VA), Italy
We propose a new procedure for quantitative evaluation of object detection algorithms. The procedure consists of a matching stage
for finding correspondences between reference and output objects, an accuracy score that is sensitive to object shapes as well as
boundary and fragmentation errors, and a ranking step for final ordering of the algorithms using multiple performance indicators.
The procedure is illustrated on a building detection task where the resulting rankings are consistent with the visual inspection of the
detection maps.
Key words: Performance evaluation, object detection, object matching, shape modeling, multi-criteria ranking
1. Introduction
Performance evaluation of pattern recognition and com-
puter vision systems has always received significant attention
(Thacker et al., 2008). Studies that characterize the theoretical
performance (Haralick, 1996; Liu et al., 2005) as well as em-


Source: Aksoy, Selim - Department of Computer Engineering, Bilkent University


Collections: Computer Technologies and Information Sciences