BCN Perceptual Computing Lab Repository

Compact Design of ECOC for Multi-class Object Categorization.

Bautista, Miguel A, and Escalera, Sergio and Baró, Xavier and Pujol, Oriol and Radeva, Petia and Vitrià, Jordi (2010) Compact Design of ECOC for Multi-class Object Categorization. Proceedings of the 5th CVCRD'10, Achievements and New Opportunities in Computer Vision . pp. 54-57.

[img]PDF (Paper) - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
330Kb
[img]Microsoft PowerPoint (Poster)
715Kb

Abstract

In this paper, we propose a Compact design of Error Correcting Output Codes (ECOC) in terms of the number of dichotomizers. Evolutionary computation is used for tuning the parameters of the classifiers and looking for the best Compact ECOC code configuration. The results over several challenging multi-class Computer Vision problems show comparable and even better results than stateof- the-art ECOC methodologies with far less cost.

Item Type:Article
Subjects:Q Science > QA Mathematics > QA75 Electronic computers. Computer science
ID Code:47
Deposited By:Xavier Baró
Deposited On:01 Dec 2010 18:43
Last Modified:01 Dec 2010 18:43

Repository Staff Only: item control page