Analysing domain shift factors between videos and images for object detection

Vicky Kalogeiton 1, 2 Vittorio Ferrari 1 Cordelia Schmid 2
1 CALVIN research group [Edinburgh]
IPAB - Institute of Perception, Action and Behaviour
2 Thoth - Apprentissage de modèles à partir de données massives
Inria Grenoble - Rhône-Alpes, LJK - Laboratoire Jean Kuntzmann
Abstract : Object detection is one of the most important challenges in computer vision. Object detectors are usually trained on bounding-boxes from still images. Recently, video has been used as an alternative source of data. Yet, for a given test domain (image or video), the performance of the detector depends on the domain it was trained on. In this paper, we examine the reasons behind this performance gap. We define and evaluate different domain shift factors: spatial location accuracy, appearance diversity, image quality and aspect distribution. We examine the impact of these factors by comparing performance before and after factoring them out. The results show that all four factors affect the performance of the detectors and their combined effect explains nearly the whole performance gap.
Document type :
Journal articles
Liste complète des métadonnées

Cited literature [42 references]  Display  Hide  Download

https://hal.inria.fr/hal-01281069
Contributor : Thoth Team <>
Submitted on : Tuesday, May 31, 2016 - 3:24:30 PM
Last modification on : Wednesday, April 11, 2018 - 1:57:57 AM
Document(s) archivé(s) le : Thursday, September 1, 2016 - 12:35:10 PM

File

vicky_pami2016.pdf
Files produced by the author(s)

Identifiers

Collections

Citation

Vicky Kalogeiton, Vittorio Ferrari, Cordelia Schmid. Analysing domain shift factors between videos and images for object detection. IEEE Transactions on Pattern Analysis and Machine Intelligence, Institute of Electrical and Electronics Engineers, 2016, 38 (11), pp.2327-2334. ⟨10.1109/TPAMI.2016.2551239⟩. ⟨hal-01281069v2⟩

Share

Metrics

Record views

698

Files downloads

396