Subjective evaluation of image understanding results

Baptiste Hemery (1) , Hélène Laurent (2) , Christophe Rosenberger (1)

1 Equipe SAFE - Laboratoire GREYC - UMR6072
2 PRISME - Laboratoire Pluridisciplinaire de Recherche en Ingénierie des Systèmes, Mécanique et Energétique

Image understanding has many applications. Given an image and a ground truth, it is possible to measure the quality of understanding results provided by different algorithms or parameters. In this paper, we ask some users to make a subjective evaluation of image understanding results by sorting them from the best to the worst. We compared the results with some provided by a metric we defined recently. Experimental results show the good behavior of this metric compared to the human judgment.

Domaines: Cryptographie et sécurité [cs.CR]

Baptiste Hemery, Hélène Laurent, Christophe Rosenberger. Subjective evaluation of image understanding results. European Signal Processing Conference (EUSIPCO), 2010, aalborg, Denmark. pp.0-0. ⟨hal-00991064⟩