Evaluating Descriptors Performances for Object Tracking on Natural Video Data

Abstract : In this paper, a new framework is presented for the quantita- tive evaluation of the performance of appearance models composed of an object descriptor and a similarity measure in the context of object track- ing. The evaluation is based on natural videos, and takes advantage of existing ground-truths from object tracking benchmarks. The proposed metrics evaluate the ability of an appearance model to discriminate an object from the clutter. This allows comparing models which may use di- verse kinds of descriptors or similarity measures in a principled manner. The performances measures can be global, but time-oriented performance evaluation is also presented. The insights that the proposed framework can bring on appearance models properties with respect to tracking are illustrated on natural video data.
Complete list of metadatas

https://hal.archives-ouvertes.fr/hal-00172137
Contributor : Rémi Megret <>
Submitted on : Friday, September 14, 2007 - 12:01:50 PM
Last modification on : Wednesday, January 31, 2018 - 1:46:02 PM

Links full text

Identifiers

Citation

Mounia Mikram, Rémi Mégret, Yannick Berthoumieu. Evaluating Descriptors Performances for Object Tracking on Natural Video Data. 9th International Conference on Advanced Concepts for Intelligent Vision Systems (ACIVS 2007), Aug 2007, Delft, Netherlands. pp.352-363, ⟨10.1007/978-3-540-74607-2_32⟩. ⟨hal-00172137⟩

Share

Metrics

Record views

251