Title :
Discriminant ECOC: a heuristic method for application dependent design of error correcting output codes
Author :
Pujol, Oriol ; Radeva, Petia ; Vitrià, Jordi
Author_Institution :
Dept. de Matematica Aplicada i Analisi, Univ. de Barcelona, Spain
fDate :
6/1/2006 12:00:00 AM
Abstract :
We present a heuristic method for learning error correcting output codes matrices based on a hierarchical partition of the class space that maximizes a discriminative criterion. To achieve this goal, the optimal codeword separation is sacrificed in favor of a maximum class discrimination in the partitions. The creation of the hierarchical partition set is performed using a binary tree. As a result, a compact matrix with high discrimination power is obtained. Our method is validated using the UCI database and applied to a real problem, the classification of traffic sign images.
Keywords :
error correction codes; matrix algebra; pattern classification; trees (mathematics); UCI database; application dependent design; binary tree; class space; compact matrix; discriminant ECOC; discriminative criterion; error correcting output code matrices; heuristic method; hierarchical partition; high discrimination power; maximum class discrimination; optimal codeword separation; traffic sign images; Binary trees; Classification tree analysis; Decoding; Error correction codes; Image databases; Mutual information; Object recognition; Support vector machine classification; Support vector machines; Visual databases; Multiple classifiers; multiclass classification; visual object recognition.; Algorithms; Artificial Intelligence; Discriminant Analysis; Image Enhancement; Image Interpretation, Computer-Assisted; Information Storage and Retrieval; Pattern Recognition, Automated;
Journal_Title :
Pattern Analysis and Machine Intelligence, IEEE Transactions on
DOI :
10.1109/TPAMI.2006.116